Project1_SysBio_NF_alldiseases_classifiers.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QPP0blfZXted"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/sayalaruano/Project3Period-SysBioMaster/blob/main/Project1_SysBio_NF_DCM_classifiers.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ihD1ikv4A_U-"
      },
      "source": [
        "# **Predicting cardiomyopathies using RNAseq data with machine learning models**\n",
        "\n",
        "This notebook is part of a group project for the Systems Biology masters at Maastricht University.\n",
        "\n",
        "You can find more details about my capstone project in this [GitHub repository](https://github.com/sayalaruano/Project3Period-SysBioMaster).\n",
        "\n",
        "## Dataset description\n",
        "We will be using an [RNA-sequencing dataset from the MAGNet consortium](https://github.com/mpmorley/MAGNet). To generate this dataset, leftventricular free-wall tissue was harvested at the time of cardiac surgery fromsubjects with heart failure undergoing transplantation and from unused donor hearts with apparently normal function.  This dataset is a spectacular resource of left ventricular transcriptomes of **non-failing healthy donors**, \n",
        "**Peripartum cardiomyopathy (PPCM)**, **Hypertrophic cardiomyopathy (HCM)** and **Dilated cardiomyopathy (DCM)** patients. Next to the RNA-sequencing data (in log2-transformed CPM values), extensive participant information is available (age, sex, ethnicity, disease state, and so on).\n",
        "\n",
        "**Note:** `mkdir`, `wget`, `unzip`, `make`, `cd`, `grep`, and `cat` are Unix based OS commands, so if you are in Windows, maybe these terminal tools won't work."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2E2XwwoUWbtw"
      },
      "source": [
        "## **Installation of libraries**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xLMn-W76a5gD"
      },
      "source": [
        "### **[LazyPredict](https://github.com/shankarpandala/lazypredict)**\n",
        "Library to run many sklearn's binary classifiers with default parameters"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DRjGMT2FVNZm",
        "outputId": "c0cf3259-dc47-4a54-e8d6-bda3a077fee0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting lazypredict\n",
            "  Downloading lazypredict-0.2.12-py2.py3-none-any.whl (12 kB)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.8/dist-packages (from lazypredict) (7.1.2)\n",
            "Requirement already satisfied: xgboost in /usr/local/lib/python3.8/dist-packages (from lazypredict) (0.90)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.8/dist-packages (from lazypredict) (4.64.1)\n",
            "Requirement already satisfied: pandas in /usr/local/lib/python3.8/dist-packages (from lazypredict) (1.3.5)\n",
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.8/dist-packages (from lazypredict) (1.0.2)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.8/dist-packages (from lazypredict) (1.2.0)\n",
            "Requirement already satisfied: lightgbm in /usr/local/lib/python3.8/dist-packages (from lazypredict) (2.2.3)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.8/dist-packages (from lightgbm->lazypredict) (1.7.3)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.8/dist-packages (from lightgbm->lazypredict) (1.21.6)\n",
            "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.8/dist-packages (from pandas->lazypredict) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.8/dist-packages (from pandas->lazypredict) (2022.7)\n",
            "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.8/dist-packages (from scikit-learn->lazypredict) (3.1.0)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.8/dist-packages (from python-dateutil>=2.7.3->pandas->lazypredict) (1.15.0)\n",
            "Installing collected packages: lazypredict\n",
            "Successfully installed lazypredict-0.2.12\n"
          ]
        }
      ],
      "source": [
        "!pip install lazypredict"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mZYPeKm97sFk"
      },
      "source": [
        "## **Imports, data obtention, and pre-processing**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 76,
      "metadata": {
        "id": "9Ql8qTq3dLSg"
      },
      "outputs": [],
      "source": [
        "# Download and unzip \n",
        "from urllib.request import urlopen\n",
        "from io import BytesIO\n",
        "from zipfile import ZipFile\n",
        "\n",
        "# Data manipulation and EDA\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "\n",
        "# Plots\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "# Machine learning classifers\n",
        "import lazypredict\n",
        "import pickle\n",
        "from lazypredict.Supervised import LazyClassifier\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.pipeline import make_pipeline\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from lightgbm import LGBMClassifier\n",
        "from xgboost import XGBClassifier\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import roc_auc_score, accuracy_score, precision_score, recall_score, f1_score, matthews_corrcoef, balanced_accuracy_score, roc_curve, auc\n",
        "from sklearn.metrics import make_scorer\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from matplotlib_venn import venn3\n",
        "from sklearn.metrics import RocCurveDisplay\n",
        "from itertools import cycle\n",
        "from sklearn.multiclass import OneVsRestClassifier\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.preprocessing import label_binarize"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "FA6Vue0LD0_t"
      },
      "outputs": [],
      "source": [
        "# Download and unzip data\n",
        "def download_and_unzip(url, extract_to='.'):\n",
        "    http_response = urlopen(url)\n",
        "    zipfile = ZipFile(BytesIO(http_response.read()))\n",
        "    zipfile.extractall(path=extract_to)\n",
        "\n",
        "url = \"https://surfdrive.surf.nl/files/index.php/s/mZrXMDlQ1Z1LvTn/download\"\n",
        "\n",
        "download_and_unzip(url, extract_to='.')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BQHwqXaV5eTe"
      },
      "source": [
        "## **Feature matrices preparation**\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8HSFMjPtd-G-"
      },
      "source": [
        "### **Load molecular descriptors (features) - training and test datasets**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "-DkYKw0yLmjL"
      },
      "outputs": [],
      "source": [
        "# Load gene expression and sampleinfo data \n",
        "gxData_cpm = pd.read_table(\"MAGNET_GeneExpressionData_CPM_19112020.txt\", index_col=0)\n",
        "\n",
        "sampleInfo = pd.read_csv(\"MAGNET_SampleData_18112022.csv\")  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "fL2qgdPYMFdS"
      },
      "outputs": [],
      "source": [
        "list_all_genes = pd.DataFrame(gxData_cpm.index)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "VxsF08pHMQYv"
      },
      "outputs": [],
      "source": [
        "list_all_genes.to_csv(\"List_all_genes.csv\", index= False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "fddqwDY69xTj"
      },
      "outputs": [],
      "source": [
        "# Transpose the gx dataframe\n",
        "gxData_cpm = gxData_cpm.transpose()\n",
        "\n",
        "# Create a new column with index values \n",
        "gxData_cpm = gxData_cpm.rename_axis('sample_name').reset_index()\n",
        "\n",
        "# Rename the index name\n",
        "gxData_cpm = gxData_cpm.rename_axis(None, axis=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "Wjr7PTapF9IE"
      },
      "outputs": [],
      "source": [
        "# Get a df with the etiology and sample name columns \n",
        "columns = [\"etiology\", \"sample_name\"]\n",
        "\n",
        "etiology = sampleInfo[columns]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "psfJU_KUF9SI"
      },
      "outputs": [],
      "source": [
        "# Join the etiology and gx dataframes\n",
        "gxData_cpm_labeled = pd.merge(gxData_cpm, etiology, on='sample_name', how='inner')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fTo2PxW9ID8U",
        "outputId": "c13d9e7c-b1d3-4909-ac08-67081e35d049"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "NF      166\n",
              "DCM     166\n",
              "HCM      28\n",
              "PPCM      6\n",
              "Name: etiology, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ],
      "source": [
        "# Prin the number of samples per etiology\n",
        "gxData_cpm_labeled[\"etiology\"].value_counts()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 552
        },
        "id": "CsSScl25AxWM",
        "outputId": "c9c73435-bc1f-44a4-98b1-98b3892d1f50"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    sample_name  ENSG00000000003  ENSG00000000005  ENSG00000000419  \\\n",
              "0        C00039            11.52             4.14            13.25   \n",
              "1        C00055            11.56             6.00            13.43   \n",
              "2        C00074             9.71             4.26            13.33   \n",
              "3        C00085            11.82             6.91            13.37   \n",
              "4        C00105            11.84             7.27            13.28   \n",
              "..          ...              ...              ...              ...   \n",
              "361      P01631            10.94            10.76            13.47   \n",
              "362      P01634            12.05             9.67            13.79   \n",
              "363      P01635            11.90             6.55            13.58   \n",
              "364      P01639            11.61             9.83            13.93   \n",
              "365      P01640            11.52             9.66            13.23   \n",
              "\n",
              "     ENSG00000000457  ENSG00000000460  ENSG00000000938  ENSG00000000971  \\\n",
              "0              11.61             9.68            11.10            14.39   \n",
              "1              11.86             9.37            10.07            14.11   \n",
              "2              11.23             9.01            13.87            12.80   \n",
              "3              11.94             9.70            11.80            14.51   \n",
              "4              11.64             9.25            11.55            14.89   \n",
              "..               ...              ...              ...              ...   \n",
              "361            10.94             8.65            10.89            14.49   \n",
              "362            11.37             8.92            10.87            14.89   \n",
              "363            11.96            10.25            11.29            14.84   \n",
              "364            11.68             9.17            10.18            14.29   \n",
              "365            10.81             8.92            10.42            14.34   \n",
              "\n",
              "     ENSG00000001036  ENSG00000001084  ...  ENSG00000273477  ENSG00000273483  \\\n",
              "0              13.40            13.23  ...             9.23             6.95   \n",
              "1              13.70            13.14  ...             8.93             5.73   \n",
              "2              13.68            13.45  ...             6.92             6.37   \n",
              "3              13.51            13.52  ...             9.35             6.64   \n",
              "4              13.90            13.14  ...             9.53             4.73   \n",
              "..               ...              ...  ...              ...              ...   \n",
              "361            13.74            12.56  ...             8.65             7.15   \n",
              "362            13.54            13.48  ...             9.02             3.59   \n",
              "363            13.44            13.50  ...             9.78             5.48   \n",
              "364            13.57            13.34  ...             9.49             6.16   \n",
              "365            13.77            12.68  ...             9.12             7.51   \n",
              "\n",
              "     ENSG00000273485  ENSG00000273486  ENSG00000273487  ENSG00000273488  \\\n",
              "0               3.40             8.78             8.69             4.99   \n",
              "1               5.18             8.96             8.02             5.18   \n",
              "2               5.48             6.58             4.26             5.84   \n",
              "3               5.48             8.98             7.85             4.59   \n",
              "4               5.87             8.88             7.16             4.73   \n",
              "..               ...              ...              ...              ...   \n",
              "361             5.96             8.57             8.84             4.83   \n",
              "362             3.59             9.28             6.76             4.33   \n",
              "363             4.34             9.93             8.56             5.72   \n",
              "364             5.77             9.68             8.48             4.39   \n",
              "365             6.47             8.89             9.10             5.55   \n",
              "\n",
              "     ENSG00000273489  ENSG00000273492  ENSG00000273493  etiology  \n",
              "0               9.56             8.76             5.28        NF  \n",
              "1               7.75             8.25             4.65        NF  \n",
              "2               9.57             5.00             5.00        NF  \n",
              "3               8.29             6.18             5.48        NF  \n",
              "4               7.27             7.16             3.99        NF  \n",
              "..               ...              ...              ...       ...  \n",
              "361             7.26             7.36             6.21       HCM  \n",
              "362             7.36             8.54             4.81       DCM  \n",
              "363             8.25             8.56             3.61        NF  \n",
              "364             8.00             8.64             5.53       DCM  \n",
              "365             6.63             8.61             7.34       DCM  \n",
              "\n",
              "[366 rows x 20783 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2494baef-5221-4773-92c5-d28debab401e\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sample_name</th>\n",
              "      <th>ENSG00000000003</th>\n",
              "      <th>ENSG00000000005</th>\n",
              "      <th>ENSG00000000419</th>\n",
              "      <th>ENSG00000000457</th>\n",
              "      <th>ENSG00000000460</th>\n",
              "      <th>ENSG00000000938</th>\n",
              "      <th>ENSG00000000971</th>\n",
              "      <th>ENSG00000001036</th>\n",
              "      <th>ENSG00000001084</th>\n",
              "      <th>...</th>\n",
              "      <th>ENSG00000273477</th>\n",
              "      <th>ENSG00000273483</th>\n",
              "      <th>ENSG00000273485</th>\n",
              "      <th>ENSG00000273486</th>\n",
              "      <th>ENSG00000273487</th>\n",
              "      <th>ENSG00000273488</th>\n",
              "      <th>ENSG00000273489</th>\n",
              "      <th>ENSG00000273492</th>\n",
              "      <th>ENSG00000273493</th>\n",
              "      <th>etiology</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>C00039</td>\n",
              "      <td>11.52</td>\n",
              "      <td>4.14</td>\n",
              "      <td>13.25</td>\n",
              "      <td>11.61</td>\n",
              "      <td>9.68</td>\n",
              "      <td>11.10</td>\n",
              "      <td>14.39</td>\n",
              "      <td>13.40</td>\n",
              "      <td>13.23</td>\n",
              "      <td>...</td>\n",
              "      <td>9.23</td>\n",
              "      <td>6.95</td>\n",
              "      <td>3.40</td>\n",
              "      <td>8.78</td>\n",
              "      <td>8.69</td>\n",
              "      <td>4.99</td>\n",
              "      <td>9.56</td>\n",
              "      <td>8.76</td>\n",
              "      <td>5.28</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>C00055</td>\n",
              "      <td>11.56</td>\n",
              "      <td>6.00</td>\n",
              "      <td>13.43</td>\n",
              "      <td>11.86</td>\n",
              "      <td>9.37</td>\n",
              "      <td>10.07</td>\n",
              "      <td>14.11</td>\n",
              "      <td>13.70</td>\n",
              "      <td>13.14</td>\n",
              "      <td>...</td>\n",
              "      <td>8.93</td>\n",
              "      <td>5.73</td>\n",
              "      <td>5.18</td>\n",
              "      <td>8.96</td>\n",
              "      <td>8.02</td>\n",
              "      <td>5.18</td>\n",
              "      <td>7.75</td>\n",
              "      <td>8.25</td>\n",
              "      <td>4.65</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>C00074</td>\n",
              "      <td>9.71</td>\n",
              "      <td>4.26</td>\n",
              "      <td>13.33</td>\n",
              "      <td>11.23</td>\n",
              "      <td>9.01</td>\n",
              "      <td>13.87</td>\n",
              "      <td>12.80</td>\n",
              "      <td>13.68</td>\n",
              "      <td>13.45</td>\n",
              "      <td>...</td>\n",
              "      <td>6.92</td>\n",
              "      <td>6.37</td>\n",
              "      <td>5.48</td>\n",
              "      <td>6.58</td>\n",
              "      <td>4.26</td>\n",
              "      <td>5.84</td>\n",
              "      <td>9.57</td>\n",
              "      <td>5.00</td>\n",
              "      <td>5.00</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>C00085</td>\n",
              "      <td>11.82</td>\n",
              "      <td>6.91</td>\n",
              "      <td>13.37</td>\n",
              "      <td>11.94</td>\n",
              "      <td>9.70</td>\n",
              "      <td>11.80</td>\n",
              "      <td>14.51</td>\n",
              "      <td>13.51</td>\n",
              "      <td>13.52</td>\n",
              "      <td>...</td>\n",
              "      <td>9.35</td>\n",
              "      <td>6.64</td>\n",
              "      <td>5.48</td>\n",
              "      <td>8.98</td>\n",
              "      <td>7.85</td>\n",
              "      <td>4.59</td>\n",
              "      <td>8.29</td>\n",
              "      <td>6.18</td>\n",
              "      <td>5.48</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>C00105</td>\n",
              "      <td>11.84</td>\n",
              "      <td>7.27</td>\n",
              "      <td>13.28</td>\n",
              "      <td>11.64</td>\n",
              "      <td>9.25</td>\n",
              "      <td>11.55</td>\n",
              "      <td>14.89</td>\n",
              "      <td>13.90</td>\n",
              "      <td>13.14</td>\n",
              "      <td>...</td>\n",
              "      <td>9.53</td>\n",
              "      <td>4.73</td>\n",
              "      <td>5.87</td>\n",
              "      <td>8.88</td>\n",
              "      <td>7.16</td>\n",
              "      <td>4.73</td>\n",
              "      <td>7.27</td>\n",
              "      <td>7.16</td>\n",
              "      <td>3.99</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>361</th>\n",
              "      <td>P01631</td>\n",
              "      <td>10.94</td>\n",
              "      <td>10.76</td>\n",
              "      <td>13.47</td>\n",
              "      <td>10.94</td>\n",
              "      <td>8.65</td>\n",
              "      <td>10.89</td>\n",
              "      <td>14.49</td>\n",
              "      <td>13.74</td>\n",
              "      <td>12.56</td>\n",
              "      <td>...</td>\n",
              "      <td>8.65</td>\n",
              "      <td>7.15</td>\n",
              "      <td>5.96</td>\n",
              "      <td>8.57</td>\n",
              "      <td>8.84</td>\n",
              "      <td>4.83</td>\n",
              "      <td>7.26</td>\n",
              "      <td>7.36</td>\n",
              "      <td>6.21</td>\n",
              "      <td>HCM</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>362</th>\n",
              "      <td>P01634</td>\n",
              "      <td>12.05</td>\n",
              "      <td>9.67</td>\n",
              "      <td>13.79</td>\n",
              "      <td>11.37</td>\n",
              "      <td>8.92</td>\n",
              "      <td>10.87</td>\n",
              "      <td>14.89</td>\n",
              "      <td>13.54</td>\n",
              "      <td>13.48</td>\n",
              "      <td>...</td>\n",
              "      <td>9.02</td>\n",
              "      <td>3.59</td>\n",
              "      <td>3.59</td>\n",
              "      <td>9.28</td>\n",
              "      <td>6.76</td>\n",
              "      <td>4.33</td>\n",
              "      <td>7.36</td>\n",
              "      <td>8.54</td>\n",
              "      <td>4.81</td>\n",
              "      <td>DCM</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>363</th>\n",
              "      <td>P01635</td>\n",
              "      <td>11.90</td>\n",
              "      <td>6.55</td>\n",
              "      <td>13.58</td>\n",
              "      <td>11.96</td>\n",
              "      <td>10.25</td>\n",
              "      <td>11.29</td>\n",
              "      <td>14.84</td>\n",
              "      <td>13.44</td>\n",
              "      <td>13.50</td>\n",
              "      <td>...</td>\n",
              "      <td>9.78</td>\n",
              "      <td>5.48</td>\n",
              "      <td>4.34</td>\n",
              "      <td>9.93</td>\n",
              "      <td>8.56</td>\n",
              "      <td>5.72</td>\n",
              "      <td>8.25</td>\n",
              "      <td>8.56</td>\n",
              "      <td>3.61</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>364</th>\n",
              "      <td>P01639</td>\n",
              "      <td>11.61</td>\n",
              "      <td>9.83</td>\n",
              "      <td>13.93</td>\n",
              "      <td>11.68</td>\n",
              "      <td>9.17</td>\n",
              "      <td>10.18</td>\n",
              "      <td>14.29</td>\n",
              "      <td>13.57</td>\n",
              "      <td>13.34</td>\n",
              "      <td>...</td>\n",
              "      <td>9.49</td>\n",
              "      <td>6.16</td>\n",
              "      <td>5.77</td>\n",
              "      <td>9.68</td>\n",
              "      <td>8.48</td>\n",
              "      <td>4.39</td>\n",
              "      <td>8.00</td>\n",
              "      <td>8.64</td>\n",
              "      <td>5.53</td>\n",
              "      <td>DCM</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>365</th>\n",
              "      <td>P01640</td>\n",
              "      <td>11.52</td>\n",
              "      <td>9.66</td>\n",
              "      <td>13.23</td>\n",
              "      <td>10.81</td>\n",
              "      <td>8.92</td>\n",
              "      <td>10.42</td>\n",
              "      <td>14.34</td>\n",
              "      <td>13.77</td>\n",
              "      <td>12.68</td>\n",
              "      <td>...</td>\n",
              "      <td>9.12</td>\n",
              "      <td>7.51</td>\n",
              "      <td>6.47</td>\n",
              "      <td>8.89</td>\n",
              "      <td>9.10</td>\n",
              "      <td>5.55</td>\n",
              "      <td>6.63</td>\n",
              "      <td>8.61</td>\n",
              "      <td>7.34</td>\n",
              "      <td>DCM</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>366 rows × 20783 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2494baef-5221-4773-92c5-d28debab401e')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2494baef-5221-4773-92c5-d28debab401e button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2494baef-5221-4773-92c5-d28debab401e');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ],
      "source": [
        "gxData_cpm_labeled"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_iDBzqjoID-p"
      },
      "outputs": [],
      "source": [
        "# Select the NF and DCM patients\n",
        "#gxData_cpm_lab_NFDCM = gxData_cpm_labeled[(gxData_cpm_labeled[\"etiology\"] == \"NF\") | \n",
        " #                                         (gxData_cpm_labeled[\"etiology\"] == \"DCM\")]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "UkHK2_iPKSux"
      },
      "outputs": [],
      "source": [
        "# Create the y label dataset\n",
        "y = gxData_cpm_labeled['etiology'].copy()\n",
        "\n",
        "# Encoding the target class label\n",
        "# y = y.astype(str).apply(LabelEncoder().fit_transform)\n",
        "y = y.map({\"NF\": 0, \"DCM\": 1, \"HCM\": 2, \"PPCM\": 3}).to_numpy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "R51pwPQLLbtN"
      },
      "outputs": [],
      "source": [
        "# Create the feature matrix by deleting the y column and the sample name\n",
        "final_df_NF_alldisease = gxData_cpm_labeled.drop(['etiology', 'sample_name'], axis=1)\n",
        "X = final_df_NF_alldisease.to_numpy()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1P1kIThkHZ_G"
      },
      "source": [
        "### **Load external dataset**"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Download GitHub repository\n",
        "!git clone https://github.com/sayalaruano/Project3Period-SysBioMaster.git"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "74003823-3cd8-4ddc-b7f5-ba61824f8098",
        "id": "yRVd13KzRj8C"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cloning into 'Project3Period-SysBioMaster'...\n",
            "remote: Enumerating objects: 186, done.\u001b[K\n",
            "remote: Counting objects: 100% (84/84), done.\u001b[K\n",
            "remote: Compressing objects: 100% (72/72), done.\u001b[K\n",
            "remote: Total 186 (delta 35), reused 40 (delta 11), pack-reused 102\u001b[K\n",
            "Receiving objects: 100% (186/186), 176.81 MiB | 16.75 MiB/s, done.\n",
            "Resolving deltas: 100% (71/71), done.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load external datasets\n",
        "ext_gxData1_cpm = pd.read_table(\"Project3Period-SysBioMaster/External_Data/Fit_data/fit_rescaled_gse116260.txt\")\n",
        "ext_gxData2_cpm = pd.read_table(\"Project3Period-SysBioMaster/External_Data/Fit_data/fit_rescaled_gse46224.txt\")\n",
        "ext_gxData3_cpm = pd.read_table(\"Project3Period-SysBioMaster/External_Data/Fit_data/fit_rescaled_gse55296.txt\")"
      ],
      "metadata": {
        "id": "7WjofPWNRj8I"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Get list of columns with DCM and NF patients\n",
        "#filter_col1 = [col for col in ext_gxData1_cpm if col.startswith('DCM')]\n",
        "#filter_col2 = [col for col in ext_gxData1_cpm if col.startswith('NF')]\n",
        "#filter_cols = filter_col1 + filter_col2\n",
        "\n",
        "# Filter the DCM and NF columns\n",
        "#ext_gxData_cpm = ext_gxData1_cpm[filter_cols]"
      ],
      "metadata": {
        "id": "x9U9dfKaRj8L"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Replace NANs by 0s\n",
        "ext_gxData1_cpm.fillna(0, inplace=True)\n",
        "ext_gxData2_cpm.fillna(0, inplace=True)\n",
        "ext_gxData3_cpm.fillna(0, inplace=True)"
      ],
      "metadata": {
        "id": "Vj1wiTwURj8O"
      },
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Transpose the gx dataframe\n",
        "ext_gxData1_cpm = ext_gxData1_cpm.transpose()\n",
        "ext_gxData2_cpm = ext_gxData2_cpm.transpose()\n",
        "ext_gxData3_cpm = ext_gxData3_cpm.transpose()\n",
        "\n",
        "# Create a new column with index values \n",
        "ext_gxData1_cpm = ext_gxData1_cpm.rename_axis('etiology').reset_index()\n",
        "ext_gxData2_cpm = ext_gxData2_cpm.rename_axis('etiology').reset_index()\n",
        "ext_gxData3_cpm = ext_gxData3_cpm.rename_axis('etiology').reset_index()\n",
        "\n",
        "# Rename the index name\n",
        "ext_gxData1_cpm = ext_gxData1_cpm.rename_axis(None, axis=1)\n",
        "ext_gxData2_cpm = ext_gxData2_cpm.rename_axis(None, axis=1)\n",
        "ext_gxData3_cpm = ext_gxData3_cpm.rename_axis(None, axis=1)\n",
        "\n",
        "# Remove numbers of etiology rows\n",
        "ext_gxData1_cpm = ext_gxData1_cpm.apply(lambda x: x.replace({'DCM.*':'DCM', 'NF.*':'NF'}, regex=True))\n",
        "ext_gxData2_cpm = ext_gxData2_cpm.apply(lambda x: x.replace({'DCM.*':'DCM', 'NF.*':'NF'}, regex=True))\n",
        "ext_gxData3_cpm = ext_gxData3_cpm.apply(lambda x: x.replace({'DCM.*':'DCM', 'NF.*':'NF'}, regex=True))"
      ],
      "metadata": {
        "id": "IVuAcV2RRj8P"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a unified external dataset\n",
        "# Create a list of the dataframe names\n",
        "pdList = [ext_gxData1_cpm, ext_gxData2_cpm, ext_gxData3_cpm]\n",
        "\n",
        "# concatenate the 3 dataframes\n",
        "ext_gxData_cpm = pd.concat(pdList)"
      ],
      "metadata": {
        "id": "Mr043hDpqfF4"
      },
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ext_gxData_cpm"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 552
        },
        "id": "H1u9uSL-rs3y",
        "outputId": "5e8e6a84-279d-4a91-9baf-2f9a52e85bd5"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   etiology  ENSG00000000003  ENSG00000000005  ENSG00000000419  \\\n",
              "0        NF            11.97             6.86            12.79   \n",
              "1        NF             4.21             7.20             4.94   \n",
              "2        NF             6.59             5.81            10.65   \n",
              "3        NF            12.14            12.75            11.10   \n",
              "4        NF             9.54             9.57            15.65   \n",
              "..      ...              ...              ...              ...   \n",
              "18      DCM             7.14            11.53             9.50   \n",
              "19      DCM             5.16            13.44             6.71   \n",
              "20      DCM             8.35             0.00            16.54   \n",
              "21      DCM            10.98             0.00             6.03   \n",
              "22      DCM            13.65            12.29            10.62   \n",
              "\n",
              "    ENSG00000000457  ENSG00000000460  ENSG00000000938  ENSG00000000971  \\\n",
              "0             11.94            10.08            11.98            15.54   \n",
              "1              3.34             3.25             2.63             3.40   \n",
              "2             12.17             3.39            13.85            12.00   \n",
              "3             11.58            13.63            10.48            12.36   \n",
              "4             13.69             9.21            11.00            11.08   \n",
              "..              ...              ...              ...              ...   \n",
              "18            11.94            11.24            12.68            14.76   \n",
              "19            12.14            11.41             2.13             9.06   \n",
              "20            10.81            12.52            10.35            17.12   \n",
              "21             9.29             3.70            10.47             7.31   \n",
              "22             9.70            16.54            15.51            12.42   \n",
              "\n",
              "    ENSG00000001036  ENSG00000001084  ...  ENSG00000273474  ENSG00000273477  \\\n",
              "0             14.00            13.57  ...             1.84             3.52   \n",
              "1             13.33            11.53  ...            13.52            13.42   \n",
              "2             10.61            12.16  ...            12.87            12.58   \n",
              "3             12.22             8.58  ...            11.72            13.19   \n",
              "4             10.66            11.49  ...             3.66             4.08   \n",
              "..              ...              ...  ...              ...              ...   \n",
              "18            15.07            11.31  ...             0.00             0.00   \n",
              "19            14.17             9.24  ...             0.00             0.00   \n",
              "20            12.43             2.05  ...             0.00             0.00   \n",
              "21            13.05            11.43  ...             0.00             0.00   \n",
              "22            15.15            14.41  ...             0.00             0.00   \n",
              "\n",
              "    ENSG00000273483  ENSG00000273485  ENSG00000273486  ENSG00000273487  \\\n",
              "0              8.30             4.87             1.40             4.20   \n",
              "1              7.86            12.58             6.33            15.97   \n",
              "2              6.86            14.20            12.19            13.33   \n",
              "3             12.98            13.80            13.51            13.28   \n",
              "4              4.04             6.38             7.77             8.82   \n",
              "..              ...              ...              ...              ...   \n",
              "18             0.00             0.00             0.00             0.00   \n",
              "19             0.00             0.00             0.00             0.00   \n",
              "20             0.00             0.00             0.00             0.00   \n",
              "21             0.00             0.00             0.00             0.00   \n",
              "22             0.00             0.00             0.00             0.00   \n",
              "\n",
              "    ENSG00000273488  ENSG00000273489  ENSG00000273492  ENSG00000273493  \n",
              "0              6.11             5.05             4.96             1.45  \n",
              "1             13.91            12.24            10.93            14.07  \n",
              "2              8.64            17.69            14.55            11.48  \n",
              "3             12.08            12.78            10.94            11.00  \n",
              "4              2.03             7.65             9.07             9.29  \n",
              "..              ...              ...              ...              ...  \n",
              "18             0.00             0.00             0.00             0.00  \n",
              "19             0.00             0.00             0.00             0.00  \n",
              "20             0.00             0.00             0.00             0.00  \n",
              "21             0.00             0.00             0.00             0.00  \n",
              "22             0.00             0.00             0.00             0.00  \n",
              "\n",
              "[90 rows x 20782 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8131a6ad-e4bc-4ec1-a57b-989737436d6d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>etiology</th>\n",
              "      <th>ENSG00000000003</th>\n",
              "      <th>ENSG00000000005</th>\n",
              "      <th>ENSG00000000419</th>\n",
              "      <th>ENSG00000000457</th>\n",
              "      <th>ENSG00000000460</th>\n",
              "      <th>ENSG00000000938</th>\n",
              "      <th>ENSG00000000971</th>\n",
              "      <th>ENSG00000001036</th>\n",
              "      <th>ENSG00000001084</th>\n",
              "      <th>...</th>\n",
              "      <th>ENSG00000273474</th>\n",
              "      <th>ENSG00000273477</th>\n",
              "      <th>ENSG00000273483</th>\n",
              "      <th>ENSG00000273485</th>\n",
              "      <th>ENSG00000273486</th>\n",
              "      <th>ENSG00000273487</th>\n",
              "      <th>ENSG00000273488</th>\n",
              "      <th>ENSG00000273489</th>\n",
              "      <th>ENSG00000273492</th>\n",
              "      <th>ENSG00000273493</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>NF</td>\n",
              "      <td>11.97</td>\n",
              "      <td>6.86</td>\n",
              "      <td>12.79</td>\n",
              "      <td>11.94</td>\n",
              "      <td>10.08</td>\n",
              "      <td>11.98</td>\n",
              "      <td>15.54</td>\n",
              "      <td>14.00</td>\n",
              "      <td>13.57</td>\n",
              "      <td>...</td>\n",
              "      <td>1.84</td>\n",
              "      <td>3.52</td>\n",
              "      <td>8.30</td>\n",
              "      <td>4.87</td>\n",
              "      <td>1.40</td>\n",
              "      <td>4.20</td>\n",
              "      <td>6.11</td>\n",
              "      <td>5.05</td>\n",
              "      <td>4.96</td>\n",
              "      <td>1.45</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>NF</td>\n",
              "      <td>4.21</td>\n",
              "      <td>7.20</td>\n",
              "      <td>4.94</td>\n",
              "      <td>3.34</td>\n",
              "      <td>3.25</td>\n",
              "      <td>2.63</td>\n",
              "      <td>3.40</td>\n",
              "      <td>13.33</td>\n",
              "      <td>11.53</td>\n",
              "      <td>...</td>\n",
              "      <td>13.52</td>\n",
              "      <td>13.42</td>\n",
              "      <td>7.86</td>\n",
              "      <td>12.58</td>\n",
              "      <td>6.33</td>\n",
              "      <td>15.97</td>\n",
              "      <td>13.91</td>\n",
              "      <td>12.24</td>\n",
              "      <td>10.93</td>\n",
              "      <td>14.07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>NF</td>\n",
              "      <td>6.59</td>\n",
              "      <td>5.81</td>\n",
              "      <td>10.65</td>\n",
              "      <td>12.17</td>\n",
              "      <td>3.39</td>\n",
              "      <td>13.85</td>\n",
              "      <td>12.00</td>\n",
              "      <td>10.61</td>\n",
              "      <td>12.16</td>\n",
              "      <td>...</td>\n",
              "      <td>12.87</td>\n",
              "      <td>12.58</td>\n",
              "      <td>6.86</td>\n",
              "      <td>14.20</td>\n",
              "      <td>12.19</td>\n",
              "      <td>13.33</td>\n",
              "      <td>8.64</td>\n",
              "      <td>17.69</td>\n",
              "      <td>14.55</td>\n",
              "      <td>11.48</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>NF</td>\n",
              "      <td>12.14</td>\n",
              "      <td>12.75</td>\n",
              "      <td>11.10</td>\n",
              "      <td>11.58</td>\n",
              "      <td>13.63</td>\n",
              "      <td>10.48</td>\n",
              "      <td>12.36</td>\n",
              "      <td>12.22</td>\n",
              "      <td>8.58</td>\n",
              "      <td>...</td>\n",
              "      <td>11.72</td>\n",
              "      <td>13.19</td>\n",
              "      <td>12.98</td>\n",
              "      <td>13.80</td>\n",
              "      <td>13.51</td>\n",
              "      <td>13.28</td>\n",
              "      <td>12.08</td>\n",
              "      <td>12.78</td>\n",
              "      <td>10.94</td>\n",
              "      <td>11.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>NF</td>\n",
              "      <td>9.54</td>\n",
              "      <td>9.57</td>\n",
              "      <td>15.65</td>\n",
              "      <td>13.69</td>\n",
              "      <td>9.21</td>\n",
              "      <td>11.00</td>\n",
              "      <td>11.08</td>\n",
              "      <td>10.66</td>\n",
              "      <td>11.49</td>\n",
              "      <td>...</td>\n",
              "      <td>3.66</td>\n",
              "      <td>4.08</td>\n",
              "      <td>4.04</td>\n",
              "      <td>6.38</td>\n",
              "      <td>7.77</td>\n",
              "      <td>8.82</td>\n",
              "      <td>2.03</td>\n",
              "      <td>7.65</td>\n",
              "      <td>9.07</td>\n",
              "      <td>9.29</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>DCM</td>\n",
              "      <td>7.14</td>\n",
              "      <td>11.53</td>\n",
              "      <td>9.50</td>\n",
              "      <td>11.94</td>\n",
              "      <td>11.24</td>\n",
              "      <td>12.68</td>\n",
              "      <td>14.76</td>\n",
              "      <td>15.07</td>\n",
              "      <td>11.31</td>\n",
              "      <td>...</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>DCM</td>\n",
              "      <td>5.16</td>\n",
              "      <td>13.44</td>\n",
              "      <td>6.71</td>\n",
              "      <td>12.14</td>\n",
              "      <td>11.41</td>\n",
              "      <td>2.13</td>\n",
              "      <td>9.06</td>\n",
              "      <td>14.17</td>\n",
              "      <td>9.24</td>\n",
              "      <td>...</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>DCM</td>\n",
              "      <td>8.35</td>\n",
              "      <td>0.00</td>\n",
              "      <td>16.54</td>\n",
              "      <td>10.81</td>\n",
              "      <td>12.52</td>\n",
              "      <td>10.35</td>\n",
              "      <td>17.12</td>\n",
              "      <td>12.43</td>\n",
              "      <td>2.05</td>\n",
              "      <td>...</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>DCM</td>\n",
              "      <td>10.98</td>\n",
              "      <td>0.00</td>\n",
              "      <td>6.03</td>\n",
              "      <td>9.29</td>\n",
              "      <td>3.70</td>\n",
              "      <td>10.47</td>\n",
              "      <td>7.31</td>\n",
              "      <td>13.05</td>\n",
              "      <td>11.43</td>\n",
              "      <td>...</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>DCM</td>\n",
              "      <td>13.65</td>\n",
              "      <td>12.29</td>\n",
              "      <td>10.62</td>\n",
              "      <td>9.70</td>\n",
              "      <td>16.54</td>\n",
              "      <td>15.51</td>\n",
              "      <td>12.42</td>\n",
              "      <td>15.15</td>\n",
              "      <td>14.41</td>\n",
              "      <td>...</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>90 rows × 20782 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8131a6ad-e4bc-4ec1-a57b-989737436d6d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-8131a6ad-e4bc-4ec1-a57b-989737436d6d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-8131a6ad-e4bc-4ec1-a57b-989737436d6d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create the y label dataset\n",
        "y_ext1 = ext_gxData1_cpm['etiology'].copy()\n",
        "y_ext2 = ext_gxData2_cpm['etiology'].copy()\n",
        "y_ext3 = ext_gxData3_cpm['etiology'].copy()\n",
        "y_ext = ext_gxData_cpm['etiology'].copy()\n",
        "\n",
        "# Encoding the target class label\n",
        "y_ext1 = y_ext1.map({\"DCM\": 1, \"NF\": 0}).to_numpy()\n",
        "y_ext2 = y_ext2.map({\"DCM\": 1, \"NF\": 0}).to_numpy()\n",
        "y_ext3 = y_ext3.map({\"DCM\": 1, \"NF\": 0}).to_numpy()\n",
        "y_ext = y_ext.map({\"DCM\": 1, \"NF\": 0}).to_numpy()"
      ],
      "metadata": {
        "id": "MTIDjU5KRj8W"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create the feature matrix by deleting the y column\n",
        "final_df_NFDCM_ext1 = ext_gxData1_cpm.drop(['etiology'], axis=1)\n",
        "final_df_NFDCM_ext2 = ext_gxData2_cpm.drop(['etiology'], axis=1)\n",
        "final_df_NFDCM_ext3 = ext_gxData3_cpm.drop(['etiology'], axis=1)\n",
        "final_df_NFDCM_ext = ext_gxData_cpm.drop(['etiology'], axis=1)\n",
        "\n",
        "X_ext1 = final_df_NFDCM_ext1.to_numpy()\n",
        "X_ext2 = final_df_NFDCM_ext2.to_numpy()\n",
        "X_ext3 = final_df_NFDCM_ext3.to_numpy()\n",
        "X_ext = final_df_NFDCM_ext.to_numpy()"
      ],
      "metadata": {
        "id": "3bhrcgfKRj8Y"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H9MRC7xR8eJi"
      },
      "source": [
        "## **Exploratory Data Analysis**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r6rVorK_rVnt"
      },
      "source": [
        "### **Target variable**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "bByC_71PtrBN",
        "outputId": "64f49e50-6d06-497a-bd0f-d6842960b550"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Bar plot of active and inactive AMPs\n",
        "\n",
        "# Data\n",
        "n_healthy = (gxData_cpm_labeled[\"etiology\"] == \"NF\").sum()\n",
        "n_DCM = (gxData_cpm_labeled[\"etiology\"] == \"DCM\").sum()\n",
        "n_HCM = (gxData_cpm_labeled[\"etiology\"] == \"HCM\").sum()\n",
        "n_PPCM = (gxData_cpm_labeled[\"etiology\"] == \"PPCM\").sum()\n",
        "\n",
        "\n",
        "x_label = ['Healthy', 'DCM', 'HCM', 'PPCM']\n",
        "y_label = [n_healthy, n_DCM, n_HCM, n_PPCM]\n",
        "\n",
        "# Setup plot\n",
        "fig, ax = plt.subplots()\n",
        "\n",
        "# Make bar plot\n",
        "p = ax.bar(x_label, y_label, color = ['#F8766D', '#00BFC4', 'red', 'green'], ec = 'black')\n",
        "\n",
        "ax.set_title('Disease status of samples', fontsize=14, fontweight='bold', pad=15)\n",
        "\n",
        "ax.set_ylim(0,200)\n",
        "plt.xticks(fontsize=14)\n",
        "plt.yticks(fontsize=14)\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RewWucDQbTo3"
      },
      "source": [
        "## **Machine learning models**\n",
        "First, we tested more than 30 ML binary classifiers using the [LazyPredict](https://github.com/shankarpandala/lazypredict) Python library. We chose the best model evaluating some performance metrics (e.g., accuracy, ROC AUC, F1 score, and Matthews Correlation Coefficient) on the test dataset. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "id": "01B14Lz2bdMn"
      },
      "outputs": [],
      "source": [
        "# Run the ML classifiers with LazyPredict     \n",
        "# Data splitting\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, \n",
        "                                                    random_state =42, stratify=y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jLk29ISs6eow",
        "outputId": "dfd8542a-6a57-499b-d24a-09e1ebf5f935"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 29/29 [06:52<00:00, 14.22s/it]\n"
          ]
        }
      ],
      "source": [
        "# Define and build the Lazyclassifier\n",
        "clf = LazyClassifier(verbose=0,ignore_warnings=True, custom_metric=matthews_corrcoef)\n",
        "\n",
        "models, predictions = clf.fit(X_train, X_test, y_train, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Delete Time column\n",
        "models.drop(['Time Taken', 'ROC AUC'], axis=1, inplace=True)"
      ],
      "metadata": {
        "id": "mJAntS6lO_vg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Jtf5g08QR48h",
        "outputId": "97295c29-0abf-4379-d08f-eec3386c2e99"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                               Accuracy  Balanced Accuracy  F1 Score  \\\n",
              "Model                                                                  \n",
              "PassiveAggressiveClassifier        0.85               0.88      0.87   \n",
              "LinearSVC                          0.82               0.87      0.85   \n",
              "SGDClassifier                      0.92               0.61      0.91   \n",
              "Perceptron                         0.76               0.59      0.81   \n",
              "LinearDiscriminantAnalysis         0.91               0.57      0.89   \n",
              "LGBMClassifier                     0.91               0.53      0.88   \n",
              "ExtraTreeClassifier                0.76               0.52      0.77   \n",
              "RandomForestClassifier             0.91               0.50      0.86   \n",
              "LogisticRegression                 0.91               0.50      0.86   \n",
              "ExtraTreesClassifier               0.91               0.50      0.86   \n",
              "CalibratedClassifierCV             0.91               0.50      0.86   \n",
              "GaussianNB                         0.72               0.50      0.73   \n",
              "XGBClassifier                      0.89               0.49      0.85   \n",
              "RidgeClassifierCV                  0.89               0.49      0.85   \n",
              "RidgeClassifier                    0.89               0.49      0.85   \n",
              "KNeighborsClassifier               0.88               0.49      0.84   \n",
              "SVC                                0.88               0.49      0.84   \n",
              "BaggingClassifier                  0.86               0.48      0.82   \n",
              "DecisionTreeClassifier             0.74               0.44      0.76   \n",
              "AdaBoostClassifier                 0.74               0.41      0.71   \n",
              "NearestCentroid                    0.62               0.38      0.68   \n",
              "BernoulliNB                        0.61               0.37      0.67   \n",
              "QuadraticDiscriminantAnalysis      0.12               0.31      0.15   \n",
              "DummyClassifier                    0.45               0.25      0.28   \n",
              "LabelSpreading                     0.46               0.25      0.29   \n",
              "LabelPropagation                   0.46               0.25      0.29   \n",
              "\n",
              "                               matthews_corrcoef  \n",
              "Model                                             \n",
              "PassiveAggressiveClassifier                 0.78  \n",
              "LinearSVC                                   0.75  \n",
              "SGDClassifier                               0.86  \n",
              "Perceptron                                  0.66  \n",
              "LinearDiscriminantAnalysis                  0.84  \n",
              "LGBMClassifier                              0.84  \n",
              "ExtraTreeClassifier                         0.59  \n",
              "RandomForestClassifier                      0.84  \n",
              "LogisticRegression                          0.84  \n",
              "ExtraTreesClassifier                        0.84  \n",
              "CalibratedClassifierCV                      0.84  \n",
              "GaussianNB                                  0.54  \n",
              "XGBClassifier                               0.82  \n",
              "RidgeClassifierCV                           0.82  \n",
              "RidgeClassifier                             0.82  \n",
              "KNeighborsClassifier                        0.80  \n",
              "SVC                                         0.80  \n",
              "BaggingClassifier                           0.77  \n",
              "DecisionTreeClassifier                      0.57  \n",
              "AdaBoostClassifier                          0.57  \n",
              "NearestCentroid                             0.44  \n",
              "BernoulliNB                                 0.43  \n",
              "QuadraticDiscriminantAnalysis              -0.13  \n",
              "DummyClassifier                             0.00  \n",
              "LabelSpreading                              0.00  \n",
              "LabelPropagation                            0.00  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-29248fca-ee8f-4047-98b4-bac624d6fccf\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Balanced Accuracy</th>\n",
              "      <th>F1 Score</th>\n",
              "      <th>matthews_corrcoef</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Model</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>PassiveAggressiveClassifier</th>\n",
              "      <td>0.85</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.87</td>\n",
              "      <td>0.78</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LinearSVC</th>\n",
              "      <td>0.82</td>\n",
              "      <td>0.87</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.75</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SGDClassifier</th>\n",
              "      <td>0.92</td>\n",
              "      <td>0.61</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.86</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Perceptron</th>\n",
              "      <td>0.76</td>\n",
              "      <td>0.59</td>\n",
              "      <td>0.81</td>\n",
              "      <td>0.66</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LinearDiscriminantAnalysis</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.57</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.84</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LGBMClassifier</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.53</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.84</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ExtraTreeClassifier</th>\n",
              "      <td>0.76</td>\n",
              "      <td>0.52</td>\n",
              "      <td>0.77</td>\n",
              "      <td>0.59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RandomForestClassifier</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.84</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LogisticRegression</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.84</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ExtraTreesClassifier</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.84</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>CalibratedClassifierCV</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.84</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>GaussianNB</th>\n",
              "      <td>0.72</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.73</td>\n",
              "      <td>0.54</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>XGBClassifier</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.82</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RidgeClassifierCV</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.82</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RidgeClassifier</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.82</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>KNeighborsClassifier</th>\n",
              "      <td>0.88</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.80</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SVC</th>\n",
              "      <td>0.88</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.80</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BaggingClassifier</th>\n",
              "      <td>0.86</td>\n",
              "      <td>0.48</td>\n",
              "      <td>0.82</td>\n",
              "      <td>0.77</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DecisionTreeClassifier</th>\n",
              "      <td>0.74</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.76</td>\n",
              "      <td>0.57</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>AdaBoostClassifier</th>\n",
              "      <td>0.74</td>\n",
              "      <td>0.41</td>\n",
              "      <td>0.71</td>\n",
              "      <td>0.57</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>NearestCentroid</th>\n",
              "      <td>0.62</td>\n",
              "      <td>0.38</td>\n",
              "      <td>0.68</td>\n",
              "      <td>0.44</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BernoulliNB</th>\n",
              "      <td>0.61</td>\n",
              "      <td>0.37</td>\n",
              "      <td>0.67</td>\n",
              "      <td>0.43</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>QuadraticDiscriminantAnalysis</th>\n",
              "      <td>0.12</td>\n",
              "      <td>0.31</td>\n",
              "      <td>0.15</td>\n",
              "      <td>-0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DummyClassifier</th>\n",
              "      <td>0.45</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.28</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LabelSpreading</th>\n",
              "      <td>0.46</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LabelPropagation</th>\n",
              "      <td>0.46</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-29248fca-ee8f-4047-98b4-bac624d6fccf')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-29248fca-ee8f-4047-98b4-bac624d6fccf button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-29248fca-ee8f-4047-98b4-bac624d6fccf');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ],
      "source": [
        "models"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Export results as a csv\n",
        "models.to_csv(\"Results_allMLclassifiers_alldiseases.csv\")"
      ],
      "metadata": {
        "id": "So0gzUN2PEd-"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 633
        },
        "id": "4ONmPLKdcKP4",
        "outputId": "7900cf07-d748-4126-eef0-83df6de6e611"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "[(0.0, 1.0)]"
            ]
          },
          "execution_count": 24,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x720 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Plot of Accuracy\n",
        "plt.figure(figsize=(5, 10))\n",
        "sns.set_theme(style=\"whitegrid\")\n",
        "ax = sns.barplot(y=models.index, x=\"Accuracy\", \n",
        "                 data=models)\n",
        "ax.set(xlim=(0, 1))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 633
        },
        "id": "Bse08KSUW1FJ",
        "outputId": "fcdcd5d7-5fd0-43e2-d178-d92b87a283eb"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "[(0.0, 1.0)]"
            ]
          },
          "execution_count": 25,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x720 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Plot of MCC\n",
        "plt.figure(figsize=(5, 10))\n",
        "sns.set_theme(style=\"whitegrid\")\n",
        "ax = sns.barplot(y=models.index, \n",
        "                 x=\"matthews_corrcoef\", data=models)\n",
        "ax.set(xlim=(0, 1))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CvITCk5anJ96"
      },
      "source": [
        "---\n",
        "We did not fine-tuned the hyperparameters of the best model because the performance metrics with the default parameters were very high, so hyperparameter tuning was not necessary."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZK0-qLC2EfYh"
      },
      "source": [
        "### **Evaluation of the best model on the test dataset**\n",
        "According to the summary tables presented above, there were some models with a perfect prediction capcity considering all of the performance metrics (values of 1 in the previous table). However, these methods are based on regression classifiers, which cannot show the feature importance of the model. So, we used the **RandomForestClassifier** that also have high performance metrics values and keep track of the feature importance (in this case, the gene expression values).\n",
        "\n",
        "Thus, we trained this model again on the entire training dataset (without creating the validation partition) and evaluated it on the **test dataset**. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kYdxIDJq3DfI",
        "outputId": "6ae2bcaa-66d6-48d4-f07a-acf7ab08ff97"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Pipeline(steps=[('standardscaler', StandardScaler()),\n",
              "                ('sgdclassifier', SGDClassifier(random_state=42))])"
            ]
          },
          "execution_count": 60,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "best_model_test = make_pipeline(StandardScaler(), SGDClassifier(random_state=42))\n",
        "best_model_test.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2KanF8QY6AV-",
        "outputId": "5fdac9d6-55ba-4090-9843-a2c44186d441"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([0, 1, 2, 3])"
            ]
          },
          "execution_count": 61,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "best_model_test.classes_"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cBmgvo_H4xH3"
      },
      "outputs": [],
      "source": [
        "# Retrieve feature names\n",
        "feature_names = pd.Series(final_df_NF_alldisease.columns, name = 'Feature')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sL6o8G555AXu",
        "outputId": "128e0998-aed6-4e06-9e2b-ba29d7b1561f"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "20781"
            ]
          },
          "execution_count": 63,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(best_model_test[1].coef_[1])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "68DfL13_f6kr"
      },
      "outputs": [],
      "source": [
        "coefs = pd.DataFrame(\n",
        "   best_model_test[1].coef_[1],\n",
        "   columns=['Coefficients'], index=feature_names\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wPsJDWYj4e4O"
      },
      "outputs": [],
      "source": [
        "coefs.plot(kind='barh', figsize=(9, 7))\n",
        "plt.title('Ridge model')\n",
        "plt.axvline(x=0, color='.5')\n",
        "plt.subplots_adjust(left=.3)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uPklLxPX20AR",
        "outputId": "ba1b8a28-fea7-4402-dd90-74d3619770d4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('standardscaler', StandardScaler()),\n",
              "                ('xgbclassifier',\n",
              "                 XGBClassifier(objective='multi:softprob', random_state=42))])"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ],
      "source": [
        "# Implementation of the best models\n",
        "best_model1 = make_pipeline(StandardScaler(), RandomForestClassifier(random_state=42))\n",
        "best_model1.fit(X_train, y_train)\n",
        "\n",
        "best_model2 = make_pipeline(StandardScaler(), LGBMClassifier(random_state=42))\n",
        "best_model2.fit(X_train, y_train)\n",
        "\n",
        "best_model3 = make_pipeline(StandardScaler(), XGBClassifier(random_state=42))\n",
        "best_model3.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "id": "NJfOehiaJ8NL"
      },
      "outputs": [],
      "source": [
        "models_names = [best_model1, best_model2, best_model3]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 104,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nmeg1qSSKB4v",
        "outputId": "7663fd8e-fc7c-4890-cbb2-3b4abe1e243d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          NF       1.00      1.00      1.00        34\n",
            "         DCM       0.82      1.00      0.90        33\n",
            "         HCM       0.00      0.00      0.00         6\n",
            "        PPCM       0.00      0.00      0.00         1\n",
            "\n",
            "    accuracy                           0.91        74\n",
            "   macro avg       0.46      0.50      0.48        74\n",
            "weighted avg       0.83      0.91      0.86        74\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "          NF       1.00      0.97      0.99        34\n",
            "         DCM       0.82      1.00      0.90        33\n",
            "         HCM       1.00      0.17      0.29         6\n",
            "        PPCM       0.00      0.00      0.00         1\n",
            "\n",
            "    accuracy                           0.91        74\n",
            "   macro avg       0.71      0.53      0.54        74\n",
            "weighted avg       0.91      0.91      0.88        74\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "          NF       1.00      0.97      0.99        34\n",
            "         DCM       0.80      1.00      0.89        33\n",
            "         HCM       0.00      0.00      0.00         6\n",
            "        PPCM       0.00      0.00      0.00         1\n",
            "\n",
            "    accuracy                           0.89        74\n",
            "   macro avg       0.45      0.49      0.47        74\n",
            "weighted avg       0.82      0.89      0.85        74\n",
            "\n"
          ]
        }
      ],
      "source": [
        "for i in models_names:  \n",
        "  # Predicting values with the best model on test and external datasets\n",
        "  y_test_pred = i.predict(X_test)\n",
        "\n",
        "  # Calculate performance metrics of test and external datasets\n",
        "  target_names = ['NF', 'DCM', 'HCM', 'PPCM']\n",
        "  print(classification_report(y_test, y_test_pred, target_names=target_names))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "id": "aYdkk_zkBi7J"
      },
      "outputs": [],
      "source": [
        "# Function to calculate performance metrics\n",
        "def performance_metrics(y, y_pred):\n",
        "  acc = accuracy_score(y, y_pred)\n",
        "\n",
        "  bal_acc = balanced_accuracy_score(y, y_pred)\n",
        "\n",
        "  prec = precision_score(y, y_pred, average = 'weighted')\n",
        "\n",
        "  recall = recall_score(y, y_pred,  average = 'weighted')\n",
        "\n",
        "  f1score = f1_score(y, y_pred,  average = 'weighted')\n",
        "\n",
        "  mcc = matthews_corrcoef(y, y_pred)\n",
        "\n",
        "  results = [acc, bal_acc, prec, recall, f1score, mcc]\n",
        "\n",
        "  return results"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OnjSbw7QBjw-"
      },
      "outputs": [],
      "source": [
        "# Dataset to store performance metrics\n",
        "results_bestMLmodel = pd.DataFrame()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "N34n43LUBm60"
      },
      "outputs": [],
      "source": [
        "for i in models:\n",
        "  # Dataset to store performance metrics\n",
        "  results_bestMLmodel_temp = pd.DataFrame()\n",
        "  \n",
        "  # Predicting values with the best model on test and external datasets\n",
        "  y_test_pred = i.predict(X_test)\n",
        "  y_ext_pred1 = i.predict(X_ext1)\n",
        "  y_ext_pred2 = i.predict(X_ext2)\n",
        "  y_ext_pred3 = i.predict(X_ext3)\n",
        "\n",
        "  # Calculate performance metrics of test and external datasets\n",
        "  results_bestMLmodel_temp[\"Performance_metric\"] = ['Accuracy', 'Balanced accuracy', \n",
        "                                                    'Precision', 'Recall', 'F1score', 'MCC']\n",
        "\n",
        "  results_bestMLmodel_temp[\"Model\"] = [type(i[1]).__name__]*len(results_bestMLmodel_temp)\n",
        "\n",
        "  results_bestMLmodel_temp[\"Test_dataset\"] = performance_metrics(y_test, y_test_pred, i, X_test)\n",
        "\n",
        "  results_bestMLmodel_temp[\"External_dataset_gse116260\"] = performance_metrics(y_ext1, y_ext_pred1, i, X_ext1)\n",
        "\n",
        "  results_bestMLmodel_temp[\"External_dataset_gse46224\"] = performance_metrics(y_ext2, y_ext_pred2, i, X_ext2)\n",
        "\n",
        "  results_bestMLmodel_temp[\"External_dataset_gse55296\"] = performance_metrics(y_ext3, y_ext_pred3, i, X_ext3)\n",
        "\n",
        "  # Append results to the general dataset \n",
        "  results_bestMLmodel = pd.concat([results_bestMLmodel, results_bestMLmodel_temp], ignore_index=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 708
        },
        "id": "foRo7sK1Bpli",
        "outputId": "3d923ccb-5f12-43cb-8e71-f6b0761e67ec"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-d68da447-7abc-42da-aa61-108045144404\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Performance_metric</th>\n",
              "      <th>Model</th>\n",
              "      <th>Test_dataset</th>\n",
              "      <th>External_dataset_gse116260</th>\n",
              "      <th>External_dataset_gse46224</th>\n",
              "      <th>External_dataset_gse55296</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Accuracy</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.63</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Accuracy</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.75</td>\n",
              "      <td>0.19</td>\n",
              "      <td>0.65</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Accuracy</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.69</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.61</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Balanced accuracy</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.61</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.53</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Balanced accuracy</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.56</td>\n",
              "      <td>0.19</td>\n",
              "      <td>0.62</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Balanced accuracy</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.49</td>\n",
              "      <td>0.56</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.57</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>F1score</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.83</td>\n",
              "      <td>0.66</td>\n",
              "      <td>0.47</td>\n",
              "      <td>0.52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>F1score</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.68</td>\n",
              "      <td>0.20</td>\n",
              "      <td>0.69</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>F1score</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.83</td>\n",
              "      <td>0.67</td>\n",
              "      <td>0.24</td>\n",
              "      <td>0.58</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>MCC</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.79</td>\n",
              "      <td>0.14</td>\n",
              "      <td>-0.52</td>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>MCC</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.82</td>\n",
              "      <td>0.22</td>\n",
              "      <td>-0.46</td>\n",
              "      <td>0.41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>MCC</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.79</td>\n",
              "      <td>0.24</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>Precision</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.79</td>\n",
              "      <td>0.71</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.54</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>Precision</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.80</td>\n",
              "      <td>0.73</td>\n",
              "      <td>0.21</td>\n",
              "      <td>0.79</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>Precision</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.79</td>\n",
              "      <td>0.66</td>\n",
              "      <td>0.23</td>\n",
              "      <td>0.61</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>Recall</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.63</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>Recall</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.75</td>\n",
              "      <td>0.19</td>\n",
              "      <td>0.65</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>Recall</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.69</td>\n",
              "      <td>0.25</td>\n",
              "      <td>0.61</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d68da447-7abc-42da-aa61-108045144404')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-d68da447-7abc-42da-aa61-108045144404 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-d68da447-7abc-42da-aa61-108045144404');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "   Performance_metric                   Model  Test_dataset  \\\n",
              "0            Accuracy  RandomForestClassifier          0.88   \n",
              "1            Accuracy           XGBClassifier          0.89   \n",
              "2            Accuracy          LGBMClassifier          0.88   \n",
              "3   Balanced accuracy  RandomForestClassifier          0.49   \n",
              "4   Balanced accuracy           XGBClassifier          0.50   \n",
              "5   Balanced accuracy          LGBMClassifier          0.49   \n",
              "6             F1score  RandomForestClassifier          0.83   \n",
              "7             F1score           XGBClassifier          0.84   \n",
              "8             F1score          LGBMClassifier          0.83   \n",
              "9                 MCC          LGBMClassifier          0.79   \n",
              "10                MCC           XGBClassifier          0.82   \n",
              "11                MCC  RandomForestClassifier          0.79   \n",
              "12          Precision  RandomForestClassifier          0.79   \n",
              "13          Precision           XGBClassifier          0.80   \n",
              "14          Precision          LGBMClassifier          0.79   \n",
              "15             Recall  RandomForestClassifier          0.88   \n",
              "16             Recall           XGBClassifier          0.89   \n",
              "17             Recall          LGBMClassifier          0.88   \n",
              "\n",
              "    External_dataset_gse116260  External_dataset_gse46224  \\\n",
              "0                         0.63                       0.44   \n",
              "1                         0.75                       0.19   \n",
              "2                         0.69                       0.25   \n",
              "3                         0.61                       0.44   \n",
              "4                         0.56                       0.19   \n",
              "5                         0.56                       0.25   \n",
              "6                         0.66                       0.47   \n",
              "7                         0.68                       0.20   \n",
              "8                         0.67                       0.24   \n",
              "9                         0.14                      -0.52   \n",
              "10                        0.22                      -0.46   \n",
              "11                        0.24                       0.00   \n",
              "12                        0.71                       0.50   \n",
              "13                        0.73                       0.21   \n",
              "14                        0.66                       0.23   \n",
              "15                        0.63                       0.44   \n",
              "16                        0.75                       0.19   \n",
              "17                        0.69                       0.25   \n",
              "\n",
              "    External_dataset_gse55296  \n",
              "0                        0.52  \n",
              "1                        0.65  \n",
              "2                        0.61  \n",
              "3                        0.53  \n",
              "4                        0.62  \n",
              "5                        0.57  \n",
              "6                        0.52  \n",
              "7                        0.69  \n",
              "8                        0.58  \n",
              "9                        0.18  \n",
              "10                       0.41  \n",
              "11                       0.06  \n",
              "12                       0.54  \n",
              "13                       0.79  \n",
              "14                       0.61  \n",
              "15                       0.52  \n",
              "16                       0.65  \n",
              "17                       0.61  "
            ]
          },
          "execution_count": 43,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Sort dataframe by performance metric\n",
        "results_bestMLmodel.sort_values(by=['Performance_metric']).reset_index(drop=True)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Dataset to store performance metrics of test dataset\n",
        "results_bestMLmodels_test = pd.DataFrame(index=range(6),columns=range(0))\n",
        "\n",
        "# Create indice list\n",
        "idxs = pd.Index(['Accuracy', 'Balanced accuracy', 'Precision', \n",
        "                                              'Recall', 'F1score', 'MCC'])\n",
        "# Set indices\n",
        "results_bestMLmodels_test.set_index(idxs, inplace=True)"
      ],
      "metadata": {
        "id": "LBtnQg2j57yq"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in models_names:  \n",
        "  # Predicting values with the best model on test and external datasets\n",
        "  y_test_pred = i.predict(X_test)\n",
        "\n",
        "  # Calculate performance metrics of test dataset\n",
        "  results_bestMLmodels_test[type(i[1]).__name__] = performance_metrics(y_test, y_test_pred)"
      ],
      "metadata": {
        "id": "_6vTtyIy572w"
      },
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "results_bestMLmodels_test"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "id": "euPsconD-EDy",
        "outputId": "765b66ac-0202-46b2-e9bf-00fa375a870f"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                   RandomForestClassifier  LGBMClassifier  XGBClassifier\n",
              "Accuracy                             0.91            0.91           0.89\n",
              "Balanced accuracy                    0.50            0.53           0.49\n",
              "Precision                            0.83            0.91           0.82\n",
              "Recall                               0.91            0.91           0.89\n",
              "F1score                              0.86            0.88           0.85\n",
              "MCC                                  0.84            0.84           0.82"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-70dce66c-555a-4fd2-a37c-0d6ccf4025e3\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>RandomForestClassifier</th>\n",
              "      <th>LGBMClassifier</th>\n",
              "      <th>XGBClassifier</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>Accuracy</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.89</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Balanced accuracy</th>\n",
              "      <td>0.50</td>\n",
              "      <td>0.53</td>\n",
              "      <td>0.49</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Precision</th>\n",
              "      <td>0.83</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.82</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Recall</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.89</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>F1score</th>\n",
              "      <td>0.86</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.85</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>MCC</th>\n",
              "      <td>0.84</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.82</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-70dce66c-555a-4fd2-a37c-0d6ccf4025e3')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-70dce66c-555a-4fd2-a37c-0d6ccf4025e3 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-70dce66c-555a-4fd2-a37c-0d6ccf4025e3');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Export results as a csv\n",
        "results_bestMLmodels_test.to_csv(\"Results_best_ensembleclassifiers_alldiseases_test_datatset.csv\")"
      ],
      "metadata": {
        "id": "TGors-Vf58Cl"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Dataset to store performance metrics of external dataset\n",
        "results_bestMLmodels_ext = pd.DataFrame(index=range(6),columns=range(0))\n",
        "\n",
        "# Create indice list\n",
        "idxs = pd.Index(['Accuracy', 'Balanced accuracy', 'Precision', \n",
        "                                              'Recall', 'F1score', 'MCC'])\n",
        "# Set indices\n",
        "results_bestMLmodels_ext.set_index(idxs, inplace=True)"
      ],
      "metadata": {
        "id": "ylVuM8nw-bD-"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in models_names:  \n",
        "  # Predicting values with the best model on test and external datasets\n",
        "  y_ext_pred = i.predict(X_ext)\n",
        "\n",
        "  # Calculate performance metrics of test dataset\n",
        "  results_bestMLmodels_ext[type(i[1]).__name__] = performance_metrics(y_ext, y_ext_pred)"
      ],
      "metadata": {
        "id": "zdt12LSx-bI6"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "results_bestMLmodels_ext"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "id": "e_6zO7dS-bN6",
        "outputId": "bc0af99f-77be-4289-9adf-751584194252"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                   RandomForestClassifier  LGBMClassifier  XGBClassifier\n",
              "Accuracy                             0.51            0.53           0.48\n",
              "Balanced accuracy                    0.60            0.50           0.43\n",
              "Precision                            0.70            0.55           0.57\n",
              "Recall                               0.51            0.53           0.48\n",
              "F1score                              0.49            0.54           0.52\n",
              "MCC                                  0.24            0.01           0.04"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-275d3a97-94a9-427e-a555-ef82c6dfb232\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>RandomForestClassifier</th>\n",
              "      <th>LGBMClassifier</th>\n",
              "      <th>XGBClassifier</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>Accuracy</th>\n",
              "      <td>0.51</td>\n",
              "      <td>0.53</td>\n",
              "      <td>0.48</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Balanced accuracy</th>\n",
              "      <td>0.60</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.43</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Precision</th>\n",
              "      <td>0.70</td>\n",
              "      <td>0.55</td>\n",
              "      <td>0.57</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Recall</th>\n",
              "      <td>0.51</td>\n",
              "      <td>0.53</td>\n",
              "      <td>0.48</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>F1score</th>\n",
              "      <td>0.49</td>\n",
              "      <td>0.54</td>\n",
              "      <td>0.52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>MCC</th>\n",
              "      <td>0.24</td>\n",
              "      <td>0.01</td>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-275d3a97-94a9-427e-a555-ef82c6dfb232')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-275d3a97-94a9-427e-a555-ef82c6dfb232 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-275d3a97-94a9-427e-a555-ef82c6dfb232');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Export results as a csv\n",
        "results_bestMLmodels_ext.to_csv(\"Results_best_ensembleclassifiers_alldiseases_ext_datatset.csv\")"
      ],
      "metadata": {
        "id": "VcN2zNWX3U3z"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## ROC curves"
      ],
      "metadata": {
        "id": "dQpnj-0lDrch"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#construct baseline pipeline\n",
        "pipe_rf = Pipeline([('clf', RandomForestClassifier(random_state=123))])\n",
        "\n",
        "# Fit the model\n",
        "model = pipe_rf.fit(X_train, y_train)\n",
        "\n",
        "#Calculate the y_score\n",
        "y_score = model.predict_proba(X_test)\n",
        "\n",
        "#Binarize the output\n",
        "y_test_bin = label_binarize(y_test, classes=[0, 1, 2, 3])\n",
        "n_classes = y_test_bin.shape[1]"
      ],
      "metadata": {
        "id": "EEhbWwuOZUgO"
      },
      "execution_count": 95,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10, 10))\n",
        "fpr = dict()\n",
        "tpr = dict()\n",
        "roc_auc = dict()\n",
        "classes = [\"NF\", \"DCM\", \"HCM\", \"PPCM\"]\n",
        "\n",
        "for i in range(n_classes):\n",
        "  fpr[i], tpr[i], _ = roc_curve(y_test_bin[:, i], y_score[:, i])\n",
        "  auc = round(roc_auc_score(y_test_bin[:, i], y_score[:, i]), 4)\n",
        "  plt.plot(fpr[i], tpr[i], lw=2, label=\"RF, \" + classes[i] + \", test_dataset, AUC=\"+str(auc))\n",
        "  # print('AUC for Class {}: {}'.format(i+1, auc(fpr[i], tpr[i])))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], color='k', lw=2, linestyle='--', label='Random guess')\n",
        "plt.xlabel('False Positive Rate')\n",
        "plt.ylabel('True Positive Rate')\n",
        "plt.legend()\n",
        "\n",
        "# Save the plot\n",
        "plt.savefig('ROC_curve_alldiseases_bestmodels.png', bbox_inches='tight')\n",
        "plt.savefig('ROC_curve_alldiseases_bestmodels.svg', bbox_inches='tight')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 606
        },
        "id": "9ekyM3eOaHr2",
        "outputId": "788e4e9c-66d9-4d4c-dc9f-1dc99639fab3"
      },
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y-EaFk3SnLvc"
      },
      "source": [
        "### **Feature importance of the best model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 105,
      "metadata": {
        "id": "ZFdeHIRXuqo_"
      },
      "outputs": [],
      "source": [
        "# Retrieve feature importance from the best models\n",
        "importance_bestmodel1 = pd.Series(best_model1[1].feature_importances_, name = 'Gini')\n",
        "importance_bestmodel2 = pd.Series(best_model2[1].feature_importances_, name = 'Importance (split)')\n",
        "importance_bestmodel3 = pd.Series(best_model3[1].feature_importances_, name = 'Importance (weight)')\n",
        "\n",
        "# Retrieve feature names\n",
        "feature_names = pd.Series(final_df_NF_alldisease.columns, name = 'Feature')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 106,
      "metadata": {
        "id": "Cq4r6KsC_tb5"
      },
      "outputs": [],
      "source": [
        "# Combine feature names and Gini values into a Dataframe\n",
        "feat_import_bestmodel1 = pd.concat([feature_names, importance_bestmodel1], axis=1, names=['Feature', 'Gini'])\n",
        "feat_import_bestmodel2 = pd.concat([feature_names, importance_bestmodel2], axis=1, names=['Feature', 'Importance (split)'])\n",
        "feat_import_bestmodel3 = pd.concat([feature_names, importance_bestmodel3], axis=1, names=['Feature', 'Importance (weight)'])\n",
        "\n",
        "# Sort the dataframe by Gini in descending order\n",
        "featimp_bestmodel1_sort = feat_import_bestmodel1.sort_values('Gini', ascending=False)\n",
        "featimp_bestmodel2_sort = feat_import_bestmodel2.sort_values('Importance (split)', ascending=False)\n",
        "featimp_bestmodel3_sort = feat_import_bestmodel3.sort_values('Importance (weight)', ascending=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 107,
      "metadata": {
        "id": "mOwz9iXF-TZB"
      },
      "outputs": [],
      "source": [
        "# Export sorted feature importance datasets\n",
        "featimp_bestmodel1_sort.to_csv(\"Important_genes_alldiseases_RandomForest_sorted.csv\")\n",
        "featimp_bestmodel2_sort.to_csv(\"Important_genes_alldiseases_LGBM_sorted.csv\")\n",
        "featimp_bestmodel3_sort.to_csv(\"Important_genes_alldiseases_XGBoost_sorted.csv\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 899
        },
        "id": "PHKPTLeIgQl3",
        "outputId": "f2747333-0839-4903-e943-a680afdfe03c"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2376x1080 with 3 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Plot of feature importance of the 3 algorithms\n",
        "#plt.figure(figsize=(5, 10))\n",
        "sns.set_theme(style=\"whitegrid\")\n",
        "fig, (ax1, ax2, ax3) = plt.subplots(1,3, figsize=(33, 15))\n",
        "\n",
        "sns.barplot(ax = ax1, x = 'Gini', y = 'Feature', data = featimp_bestmodel1_sort.iloc[0:19])\n",
        "ax1.title.set_text('RandomForest')\n",
        "sns.barplot(ax = ax2, x = 'Importance (split)', y = 'Feature', data = featimp_bestmodel2_sort.iloc[0:19])\n",
        "ax2.title.set_text('LGBMClassifier')\n",
        "sns.barplot(ax = ax3, x = 'Importance (weight)', y = 'Feature', data = featimp_bestmodel3_sort.iloc[0:19])\n",
        "ax3.title.set_text('XGBClassifier')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SiQd88ltDLcf"
      },
      "source": [
        "The feature importance results show that most of the genes have low Gini impurity values, which means that all elements belong to a certain class and there are not highly important genes that helps to the classification task."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 108,
      "metadata": {
        "id": "m0DeCke81Ojy"
      },
      "outputs": [],
      "source": [
        "# Get the top 50 most important genes for each method\n",
        "imp_genes_RF = featimp_bestmodel1_sort.iloc[0:500]\n",
        "imp_genes_LGBM = featimp_bestmodel2_sort.iloc[0:500]\n",
        "imp_genes_XGB = featimp_bestmodel3_sort.iloc[0:500]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 109,
      "metadata": {
        "id": "gzB_KZDD1QDC"
      },
      "outputs": [],
      "source": [
        "# Obtain the top 50 most important genes that are present in at least two methods\n",
        "consensus_genes1 = imp_genes_RF.merge(imp_genes_LGBM,on='Feature')\n",
        "consensus_genes2 = imp_genes_RF.merge(imp_genes_XGB,on='Feature')\n",
        "consensus_genes3 = imp_genes_XGB.merge(imp_genes_LGBM,on='Feature')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 110,
      "metadata": {
        "id": "c6VqBBO71QMB"
      },
      "outputs": [],
      "source": [
        "# Concatenate the results\n",
        "df_final = pd.concat([consensus_genes1, consensus_genes2, consensus_genes3], axis=0, ignore_index=True)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Remove duplicates \n",
        "df_final.drop_duplicates(\"Feature\", inplace=True)"
      ],
      "metadata": {
        "id": "8B5lgppUgaua"
      },
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 112,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 488
        },
        "id": "R0qiDNIA-UK4",
        "outputId": "50908341-e128-4e4d-a567-d5dbd38fb720"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "             Feature  Gini  Importance (split)  Importance (weight)\n",
              "0    ENSG00000205795  0.02               28.00                  NaN\n",
              "1    ENSG00000162998  0.01               12.00                  NaN\n",
              "2    ENSG00000075413  0.01                7.00                  NaN\n",
              "3    ENSG00000139289  0.01                4.00                  NaN\n",
              "4    ENSG00000123243  0.01                3.00                  NaN\n",
              "..               ...   ...                 ...                  ...\n",
              "212  ENSG00000079931   NaN                3.00                 0.00\n",
              "213  ENSG00000065978   NaN                2.00                 0.00\n",
              "214  ENSG00000115290   NaN                3.00                 0.00\n",
              "215  ENSG00000197299   NaN                5.00                 0.00\n",
              "216  ENSG00000196843   NaN                2.00                 0.00\n",
              "\n",
              "[161 rows x 4 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-0bb11362-acb5-4a37-ab7c-d3d329841189\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Feature</th>\n",
              "      <th>Gini</th>\n",
              "      <th>Importance (split)</th>\n",
              "      <th>Importance (weight)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ENSG00000205795</td>\n",
              "      <td>0.02</td>\n",
              "      <td>28.00</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ENSG00000162998</td>\n",
              "      <td>0.01</td>\n",
              "      <td>12.00</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>ENSG00000075413</td>\n",
              "      <td>0.01</td>\n",
              "      <td>7.00</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>ENSG00000139289</td>\n",
              "      <td>0.01</td>\n",
              "      <td>4.00</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>ENSG00000123243</td>\n",
              "      <td>0.01</td>\n",
              "      <td>3.00</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>212</th>\n",
              "      <td>ENSG00000079931</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>213</th>\n",
              "      <td>ENSG00000065978</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>214</th>\n",
              "      <td>ENSG00000115290</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>215</th>\n",
              "      <td>ENSG00000197299</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>216</th>\n",
              "      <td>ENSG00000196843</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2.00</td>\n",
              "      <td>0.00</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>161 rows × 4 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-0bb11362-acb5-4a37-ab7c-d3d329841189')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-0bb11362-acb5-4a37-ab7c-d3d329841189 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-0bb11362-acb5-4a37-ab7c-d3d329841189');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 112
        }
      ],
      "source": [
        "df_final"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 113,
      "metadata": {
        "id": "SZwj60gs3hF1"
      },
      "outputs": [],
      "source": [
        "# Export consensus genes\n",
        "df_final.to_csv(\"Consensus3methods_500important_genes_alldiseases.csv\")"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Ven diagram"
      ],
      "metadata": {
        "id": "e-cdyDW8W3Kc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "imp_genes_RF_set = set(imp_genes_RF['Feature'])\n",
        "imp_genes_XGB_set = set(imp_genes_XGB['Feature'])\n",
        "imp_genes_LGBM_set = set(imp_genes_LGBM['Feature'])"
      ],
      "metadata": {
        "id": "20w-qTtTW2gb"
      },
      "execution_count": 114,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create Ven diagram\n",
        "plt.figure(figsize=(10, 10))\n",
        "venn3([imp_genes_RF_set, imp_genes_XGB_set, imp_genes_LGBM_set], ('Random Forest', 'XGBoost', 'LightGBM'))\n",
        "\n",
        "# Save the plot\n",
        "plt.savefig('Ven_diagram_alldiseases_bestmodels.svg', bbox_inches='tight')\n",
        "plt.savefig('Ven_diagram_alldiseases_bestmodels.png', bbox_inches='tight')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 548
        },
        "id": "5P_DE4mOW2kz",
        "outputId": "f77df575-5dab-4c5a-e1e7-264edca87104"
      },
      "execution_count": 118,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjwAAAITCAYAAAAD9cZ5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd5xcd33v/9d3+s527a56l2VZkuXecQNTDIQSIJiAMZhLC6RACCW5zt1sAiHhBocklEvg3gRC+YUASag2LtjGxlUukq1iWb2t2vadPvP9/XFm8VpW2d0pZ8457+fjMQ+tdlcznx3tOed9vtVYaxERERHxs5DbBYiIiIjUmgKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J4Cj4iIiPieAo+IiIj4ngKPiIiI+J6vAo8x5i+MMd9yuw4RERFpLDUPPMaYXcaYtDFmzBjTb4z5V2NMS61ft5aMMdcaY0rln2ni8eM6vv5SY4w1xkTq9ZoiIkFmjGkpX8/eMelzrcaYPcaYt5T/fpEx5ifGmEFjzJAxZpMx5jPGmM7y199tjClOum7sMMb8Xo3rvtYYs6+Wr+EV9WrheZ21tgU4Dzgf+NM6vW4tHbDWtkx6vG66T2CMCdeiMBERqS5r7RjwAeALxpie8qc/Bzxmrf2+MeYK4B7gAeAsa20HcD1QAM6d9FQPTlw3gDcDnzPGnF+vnyPI6tqlZa3tB27HCT4AGGM+ZYzZbowZLafh3570tXcbY+43xvxdOTHvNMa8etLXlxlj7i3/2zuA7smvZ4x5vTHmmXLSvscYs3rS13YZYz5ujNlgjBk3xvxfY8wcY8zPy89350Qqnw5jzOryaw2VX/v1k772r8aYrxhjfmaMGQdeaoyZb4z5gTHmSPnn+8NJ33+JMeYxY8yIMeaQMebW8pfuK/85VL5LuHy6dYqIyPRYa28Hfgr8ozHmWuCtwIfKX/4c8C/W2s9aaw+Vv3+PtbbXWnvPSZ7vCWAzMPnadKrr1qmuL68pX0NHjTH7jTF/YoxpBn4OzJ/UqjS/mu+Jp1hra/oAdgEvL3+8ENgI/MOkr/8OMB8nfN0AjAPzyl97N5AH3geEgd8DDgCm/PUHgVuBOHA1MAp8q/y1M8vP9QogCnwCeA6ITarrIWAOsAA4DDyO0wKVAO4Gek/yM10L7DvB56Pl1/gzIAa8rFzTqvLX/xUYBl5S/nmTwHrgf5W/fzmwA3jVpJ/vneWPW4DLyh8vBSwQqfX/nx566KGHHs8/gE7gIHAUuLn8uWagCFx7mn/7buD+SX+/GBgCziz//aTXrSlcXw4CV02q8YLyxye8XgXxUa8Wnv8yxowCe3GCRe/EF6y1/2GtPWCtLVlr/x3YBlwy6d/uttZ+zVpbBL4BzAPmGGMW4/yy/Lm1NmutvQ+YPI7mBuCn1to7rLV54O+AJuCKSd/zT9baQ9ba/cCvgIettU9YazPAf+KEn5OZX07ZE4+3ApfhBJO/sdbmrLV3Az8BfnfSv/tva+0D1toSsA7osdb+Zfn7dwBfA95W/t48cIYxpttaO2atfeiU77KIiNSUtXYQeAbnhvWH5U934tzE9k98nzHmc+Vrw7gx5pZJT3FZ+fOjwCPAv+Fc9+DU163TXV/ywBpjTJu1dtBa+3j1f3pvq1fgeaO1thUnaZ7FpK4nY8xNxpgnJ4IDcDYv7Jr6zS+QtTZV/rAFp1Vo0Fo7Pul7d0/6eP7kv5cDxl6c1pwJhyZ9nD7B3081uPqAtbZj0uN75dfcW36tyTVNfs29kz5ewnHBCSe9zyl//X/gJP4txphHjTG/dYp6RESkxowxN+K0st8J/G3504NACeeGHABr7SesM47nP4HJE0weKl8zWoG5wFrgr8tfO9V163TXlzcDrwF2l4d6aKjDceo9hudenG6dvwMwxizBadH4faCr/MvxNGCm8HQHgc5yH+WExZM+PoATKCi/lgEWAfsr+BFO5wCwyBgz+X1dfNxr2kkf7wV2HhecWq21rwGw1m6z1v4uMBvnwPp++eed/BwiIlIHxpjZwN/jDLP4APBWY8xV5Rvvh4E3Tef5rDPW5wfAxKSXU123Tnl9sdY+aq19A8714r+A7028zHRq8jM31uH5AvAKY8y5OP2eFjgCYIy5GaeF57SstbuBx4A+Y0zMGHMlz//SgPOf/VpjzHXGmCjwMSAL/LpqP8mLPQykgE8YY6LlQW2vA/6/k3z/I8CoMeaTxpgmY0zYGHO2MeZicO4kjDE95UQ/VP43JZz3q4Qz5kdEROrji8B/WWt/aa09iDPG5mvGmHj54/cYZyLObABjzEJg2cmezBjTBfw2ThcZnPq6ddLrS/ka+A5jTHu5K2wE5xoBTs9FlzGmvYrvgyfVfR0Xa+0RY8w3gf9lrX2zMebzOINzS8A3cab0TdXbccb1DJSf45tAR/l1tpabHv8Jp8nvSZzp8bmq/TDHsdbmjDGvA76MM/V+P3CTtXbLSb6/WO6m+jywE2fw9VZgor/3euBWY0wSp+nybdbaNIAx5jPAA+WD4vpAje9x7nrCkx6R4/5ucAYQFnGmhBZf8Hhhk7BIYJk+E8E570RwjpuTPUo4x1IBZ6xI3vYG6zgyxrwRuBJYM/E5a+3XjTFvx7me/U9jzMtwxqh+yjlNsQ/4b5zr0ITLjTFj5Y9TwF3AH5Wf75TXrZNdX4wxMeCdwBeNs9zJVuAd5efcYoz5LrCj/LU11toDVX57PGFitpOI+5yDth2n5W/ikcQZtJeY9GelQb2EM0YrjXPCOf7jFDDM82PGRDzD9JkQ0HrcI4kTbI5/VLIWWAHIlR8ZYAxnhtHkxxiQtr260Ij7FHikvpzbnhaclrjJj3ack3IjyeN0JQ6X/xz6zd+dWYMirjF9pgnowpnk0ckLw81UxkHWSwmni2Ww/Bgo/zkctFYicZcCj9SWM8h6Ds5Autk4J2evb4lhcYLPYZz+8cPAIDqYpEZMn2nDOXa6cUJOF413gzBdJZybiQGeP5aOKgRJrSjwSPU4swcmgs1EyGk+5b/xjzzOYPKJE/dBajheTPzN9JkOnGnI88oPr4ebqSrgHEf9OMfRIdtrs+6WJH6hwCOVcTaCXVR+zMdZAVScVqDDOEsP7MXaIy7XIw3M9Jl2nl9rZR7OeDVxDAB7yo9DGg8kM6XAI9PjtOLMxVn/YRHO2AE5vTTOjI29wD6c1bwlwEyfmY2zgN1SyrNL5bSyOMfRHmCv7dVxJFOnwCOn54SchcAZOItiRd0tyPMsTpP9c8AOrJrsg6A8e2o+TsBZQnC6e2tlYg23HcBztlezKuXUFHjk5IyZixNyluNMB5fqK+HcsT4H7MLagsv1SJWZPjMHWIVzHKnLtzYszur724Cdtlfj5+TFFHjkhYyZBawEVnDqvcSk+grALpzws1ezvrzL9Jkkzj54Z6Luqnor4izUug2n20uzvgRQ4BGY6LJahrOJ3VyXqxHHOLAF2KwFEL3B9JkwTlfVKpwu4EZaCyeoMsBmYJPtfcFG0xJACjxB5mxZsbr8CMq0V68p4dytPo2zd480mHJrzhqc40izqxpTCaf19Bnbq+MoqBR4gsgZm7MWp1XHjQ1kZWaOARuB7Vrp2X2mz3QD63C6f3UceccA8DTOQGeNmQsQBZ4gMWYxcAHOgoDiXSlgA7BJg5zrz/SZhcC5OOvmiHdlcW4gntYg52BQ4AkCJ+hcCPS4XYpUVRp4CgWfujB9ZjFwEc72DuIfOZzgs1HBx98UePzMmCU4QUcnaH/L4ASfZxR8qs/0mfnAxTjbpYh/5XC6ujZqOwt/UuDxIwWdoMrgdHU9reBTufJKyBejrqugmQg+G9Ti4y8KPH5iTBdwOc5qrhJc48DDWPuc24V4kekzs3CCzhK3axFXZYD1wGat5eMPCjx+YEwC5wR9Flr7Q553CPi1Ni6dGtNn4jjH0Wp0HMnzhoCHbK/d43YhUhkFHi9zFgw8G2fmlZasl5N5FnhECxiemOkzBmexwEvQFipycnuAB22vHXa7EJkZBR6vcmZeXQ60u12KeEIBeALYoDV8nmf6TA9wJZrBKFNTwpnRtV5r+HiPAo/XOKsjvwRn0UCR6RoC7sXaQ24X4ibTZxI4LTpnuV2LeNIIcJ/ttQfcLkSmToHHS4xZBVwGxN0uRTzN4tylPhbE2VymzywDrkLdV1K5LTjjezSbywMUeLzAmBbgapwNCUWqZRintaff7ULqoTwo+SXAGW7XIr4yDvxKg5obnwJPo3NadS5Hg5Kldp7GGdTs29Ye02cWAdegTXKldp4Dfm17bcbtQuTEFHgalTPV/FpgscuVSDCMAL/029ge02eiODcMGqsj9ZAC7tbYnsakwNOInN3MrwOa3S5FAqUEPIq1T7ldSDWUV0q+Dmh1uxYJFIszI3K97dUFtpEo8DQaY87D2aAw5HYpElh7cFp7PLufkOkzZ+MM8NdxJG45CNxle7X+VaNQ4GkU6sKSxjIG3Im1h90uZDrKXVhXAyvcrkUEZ3uKezSguTEo8DQCY+bgNL23uF2KyCQlnMHMG9wuZCpMn+kEXgF0uF2LyHGeAh5RF5e7FHjcZsw64FLU9C6NazdwN9bm3S7kZEyfOQOnZSfidi0iJ7EPuFNr9rhHgcctzj5YVwBr3C5FZAqOAbdh7bjbhUxW3gfrMmCd27WITMEQcJvttSNuFxJECjxuMCYGvBwtJCjeksIJPUfdLgTA9JkI8DJgqculiExHFrhDU9frT4Gn3pxVk18NdLpdisgMFHC6t3a5WYTpM03Aq4DZbtYhMkMl4AHbaze7XUiQKPDUkzGzcU7STW6XIlIBCzzs1mBm02c6cG4atL6OeN3TwIMazFwfCjz1YsxynGnnGlQpfrEZuJ86nkRMn5kPvBJttSL+sR34pe21JbcL8TsFnnow5iyc3ZmN26WIVNlO4C5s7U/Wps+sAF6KZjSK/+zBGddTdLsQP1PgqTVj1uLs0CziV3uAO7C1O1mbPnMmzuafumkQvzqIM4OrYZd/8DoFnloy5lycNXZE/G4f8Ita7Lhu+tRCKoFxBPi5dlyvDQWeWjHmApw9sUSC4iDOtPWq3aGaPrMGuLJazyfiAYPAT7UHV/Up8NSCMZcA57ldhogLDgM/w1a+mqzpM+uAyysvScRzRoAfKfRUlwb/VZsxl6GwI8E1G/it8uKaM2b6zLko7EhwtQGvNX0m4XYhfqLAU01ON9Y5bpch4rJu4HqMmdESDKbPrEVj30Q6gVebvspuHuR5CjzVYswaNGZHZMJc4OXlPeOmrDz1XLMaRRw9wPXlbVSkQgo81eAsKqiTtMgLLcaZSj4lps8sxFlnR0SeNxd4pemb3s2DvJjewEoZswBnA0NNmRV5sZUYc9qxOKbP9ACvQOckkRNZCFxn+oyuMxXQyaUSxvTgLHOv91Hk5NZhzEkH8k/aGytav5JEPGcZcIXbRXiZLtQzZUw7OkmLTNUl5S1WXsD0mSTwGkCzUUROb215UL/MgALPTDhTbl+FTtIi03Elxsyf+Et5IOb1QIt7JYl4zhXl8W4yTQo802WMwRmz0+F2KSIeE8KZudVW/vu1OFPYRWTqDPDyclewTINWWp4uraIsUqnBrk+yc6CJC9wuRMTDRoD/0r5bU6cWnulwpp8r7IhU4MAiEn9yhNVu1yHicW1ouvq06I2aKmO6cJrgRWSGxlrJPXkZsXVpmm4+zLjb9Yh43Fy0BcuUKfBMhTEJnOnnWu1SZIYKEUoPvRRKEcIAbxwkefkoabfrEvG4tabPLHO7CC9Q4JmalwGtbhch4mXrX0Im08xv9gUKgfnjg8S68xTcrEvEB64xfUbXqNNQ4DkdY9bhrHIpIjO0ZzmpI/NJHv/5hCV8y34KxqLZEyIzF8OZuaVr+inozTkVY2YBl7hdhoiXjTeTf/pC4if7+oosiRuPkqpnTSI+1ANc6nYRjUyB52SMCQPXgTPeQESmz4J99BqKE+N2TubNAyRXpcnWqy4Rn1pn+swSt4toVAo8J3cZ0Ol2ESJetvk80mPtp1+RPAzmT/dj4iVK9ahLxMeuNX2m2e0iGpECz4kYsxjQfiUiFTjWQ2bHWTRN9fu7isQ+dhAtoiZSmThwldtFNCIFnuMZ0wRc43YZIl5WiFBafyVhDGY6/+7yMZLXjGiqukiFFps+s8rtIhqNAs+LvQSmflcqIi+28SIyuQTRmfzbDx4imixSrHZNIgFzubq2XkiBZzJjFgHL3S5DxMsGusnuXzrzm4aWEpEPHdIAZpEKxVDX1gso8EwwJgJc6XYZIl5WMtgnL4fpdmUd75pRkutSGs8jUqHFps+scLuIRqHA87wL0GrKIhXZtpZUquXka+5Mx0cOEgprQUKRSr3E9JnTzpQMAgUemFhg8By3yxDxslSS/HNrqjf+bXaB2E1HtCChSIUSaAFdQIFnwlXovRCpyBNXULDh6h5Hrx+kaUGWXDWfUySAVpk+0+V2EW7TRd6Y1cAct8sQ8bJ9S0kN9lR/dmMEQn/crxlbIhUywBVuF+G2YAceY2LAxW6XIeJlJYPddP7MpqBPxZkZmi4e09o8IhWaZ/pMoGchBzvwOAOVNZhLpALPrSE10zV3puoDhwhrR3WRil1q+kxg94cMbuAxphVtHyFSkXyU4vbVtb9pmFMg9pohtfKIVKiVAE/QCW7gcUatBzbpilTDlnPJFKP1OY7ecZSYNhcVqdj5ps8k3S7CDcEMPMZ0A1qMSaQC6SYKe5bXbxuW1hKRtx9VK49IhSLAhW4X4YZgBh4NVBap2DMXkqv2NPTT+a0hmjoKFOr5miI+tMr0mRa3i6i34AUeY+YBi9wuQ8TLhjvI9S+s/ya7MUvopiPaZ0ukQiHgfLeLqLfgBR64yO0CRLxu8/kUKt0va6auHaGpraC1eUQqFLhWnmAFHmNmA/PcLkPEy8ZbyB+dU//WnQlRCL3tmFp5RCoUuFaeYAUeOM/tAkS8bss55N1q3ZnwymHiyaJaeUQqFKhWnuAEHmM6gKVulyHiZZkmCv2L3F+sM24Jv3mAjNt1iHhcoFp5ghN4ArzYkki1PLuWnA01xnnjtUPEtS6PSMVWmT7jWhd1PTXEiavmjEkCZ7pdhoiX5aMU9y0j7nYdE5pLRF43qFYekQqFgNVuF1EPwQg8sI7g/KwiNbF9NZlSpLFWJ3/DIFHtsSVSsTWmz/j+Gun7H7C8I3og0qtIrZQMdtfKxmndmdBRJHrZmFp5RCqUBHy/k7r/Aw+sBGJuFyHiZQcXkynEiLhdx4m8ccDtCkR84Wy3C6i1IASes9wuQMTrdq5yu4KTOytDoidP3u06RDxutukzPW4XUUv+DjzG9ABdbpch4mXjzeSHZrk/Ff1kQmDeNEDO7TpEfMDXrTz+Djxq3RGp2M5V5NxeaPB0rh0hHtbgZZFKrTB9pmFvbirl38BjTARY4XYZIl5mwTbSVPSTaSkRuXqEtNt1iHhcCB8PXvZv4HH+0zRYWaQCBxeTbdTBysd73aCvz2ci9bLS7QJqxc8nCE1FF6nQzjO90020Ikt8VoGC23WIeNwc02fa3C6iFvwZeJx9s+a4XYaIl2XjFAe7Gnew8vFCYK4b1uBlkSo4w+0CasGfgcfHfZAi9XJgCVlCjT1Y+XhXjzTWStAiHuXLbi1P9M3PwDK3CxDxuv1LvHdDtDRHvCdP/kiUqNu1SINIEeHLfJwSESxhFrCeG/kxt/JxCuUWzDyttLGTP+ArHCXJd3kXaXoIked6vsHZHHD5p6i3dtNnZttee9jtQqrJf4HHmFa09o5IRXIxikOzGn921om8fJjcd7sVeKQsQYH3cSvtZMkS5p/4OI/yNH/M//7N93yRD7KUJwH4Ca+mk738AV9hM3O5jd/lbP7erfJdtBLwVeDx3B3cFKh1R6RCBxZ7rztrwjUjPryRk5kLAe1kAcgTxh7X7TlIgkFW8ZJy4BlhHivYAsBq+snQxQFa61pzY1jqdgHVpsAjIi+yf6k3ww7AgjzxeTltNSGTFDB8jj/nVv6OHjZzMTt/87UHOI9OttBZ3oS2g31s5gIAHmMpWbrop9OVut3VbPqMr3pL/BV4jEmi2VkiFclHvTU760SuHdFsLZkkguUT/BUf5pMMspSnmf+br23nEs7kkd/8/XXcRp4mPsef8ygvo4W9hCm5UXYDWOJ2AdXkt6bfwLTuHIXIOvh4ASIlCF8M62+DH/8trLoV3lKEyCLY/QB8MwmlrZB8A7zrGPREIP/38I23EbiBeDIFBxeRJUTS7ToqcdE4oe92u12FNJwu0sxmK8+wlrM5wCFaGGEpV/Ll33xPJxk+wDcAKAF/w1+zmKMuVey2xcDjbhdRLf5q4fFZGj2VTig8Arcegb/aBX+1Adb+Eyz/NNz8NfjaUeibAwN/AJcDfBBevRz2HoG//DL8y6fgBpd/BGlQ/QvdrqByyzLEo6XA3pXLZP20cIwmAMaJcog19NAPwK+5gB42kJy0YOUxmsiUx/n8iCuZxbbfdHcFT4+f9tbyTwuPMWFgnttl1EsYWIQzEG8cwkUIR8GGofD68sj6V8GmL8GrgQf2wbwPwW0Avw39N0PXemi9EEbd+ymkEQ30eH9LliiEzkuRfrSlfKGT4DpMO7dzM5YQFsNCHuNlbARgFxdzvnNe/I3tzONubgYsLRzgbXzThaobhcFp5XnW7UKqwT+Bxxm7E6hFxzJgFsMtg9BzFdzzPtj5CQh/HZa8F3b/AC4cwhlstwT2/Sdc8FF47quwdAS6noROBR6ZbLSNXCHm/cADcNkYpUdb3K5CXHcO+zmHT5/wax/l8y/63CXs4BL+vNZleYhvAo+furTmn/5b/CUB9jD81Sb45HZY+h8w/7PwtT546zz40yRkQjh7IX0NbhuDptnw51+Bl82BvVHU5C8vdGiBf/aiOm9ca/GIVMFC02c8O2tzMj+18AQu8ExYCemzYet/wNofwB0fxllQ69OwZn951toyyDyOMxCvCHTCX19FYAfiyUkcnufd6ejHm10g1p2ncDTqq/OcSL3FcBbz9fz1wh8tPMZEgNlul1FPT0HLNpzxCYchugHWrIb+J3AWyBqEyNfhVTfBvQDboGm43OX3XrhyOWxbRmAH4skJWLDDXf7ozppw6Zimp4tUgS+We/HLnc9c/BLepuhpaP8o3GwhZMFcCo99Gja+FN68Ac6xYF4B934StgL8AubdAjcbsHPgwH8R6IF4cgJDXeSKEW9uJ3Ey56YwPw3iknEi1TUXeMbtIiplrLVu11A5Yy4BznO7DBEv27qO8W1n0+x2HdV0NELu5hX+arUSccG47bXfdruISvmlVSQw09FFauVYj2/OB7/RXSDWVqDodh0iHtds+ozn5zx6/wRnjEG7o4tUbLTDny0ha9MaxyNSBXPdLqBS3g880IF/xiKJuCLTRCEf9+c6VmvT+KDfXsR1CjwNQK07IhU61uPf3cVXpv0z1V7ERZ6fCe2HwKMtAkUqNNTl31aQJTktQChSBR1eX4BQgUdEGJ7li3PBCTWXiHTn/bOCtIhLIkCb20VUwg8nOXVpiVRorM2f43cmLMkq8IhUgadXtfJ24DGmBfy1UJpIveWjFHMJf3f7LM1q3ziRKpjldgGV8PrsJrXuiFRotJ0C1KaFZyRL5Pd/xseLloi1hFd2sb7vWn78kdu46UiKJYBpj3PoL1/Kv3YnydaiBoAlmpguUg0KPC5qd7sAEa8ba6vdwnwtMQq3vopbu5Nk03nCH/gJH//ZNp7+0yv53pwWZy+3T93J7/yfx3jpLVdzW63qWJDTTC2RKlCXlota3S5AxOvG2mo3QytkYKLlJlMgXLKEDTARdkoW8iWihtrOEpuT9/zNnUgjaDd9xrO5wesnAQUekQqNt9b2xidXxLznv7llLEfP2h7uefVKdgJ85DbetXeEde1xDtxyFd+vZQ3tRaLxEqVsyPM3eSJuCuFcd4fdLmQmvH7wK/CIVCjdXNvunlgY+6038Vdffi2fPDjG0vt2Mx/gC9fzje++mY/PaqL/Wxu5qJY1ACzWTC2RavDsdVeBRyTg0sn6tPTObyW9pJ2tD+xh7cTnYmHslYt5dOMhLqj16y/MKfCIVIFnNxH1buAxpgnvd8mJuKoYopSP1+442jlIy4FRmgCGM0R3DrFmYRuHHj1ADzhjeB7ez7ldSfprVcOEbsUdkWrwbEODlwODZ990kUaRaqEAtdslffcw7V9/nJsthKzFrOrisd9dx8b3/oiP54skLJiuJPtuuYpv16qGCV0KPCLV0Ox2ATPl5cDj6SWuRRpBurm2C/Jdu5T91y7l08d//l/fyOdq+bonMkuBR6Qakm4XMFPe7dKChNsFiHhdLh6cFYg7Cp4+34k0Cs+28Hj5BKAtJUQqlA3QbUN7UYsPilSBWnhcoMAjUqFcvLYL/jWStqK/N0gVqZO46TOevHlQ4BEJsCC18CRLhI0NTsATqSFPbjaswCMSYPlYcLp5QmCaS8EZsyRSQwo8dVazqbQiQZGLByfwACRKauERqQIFnjpTC49IhfIxT58Dpi2qLi2RavBkg4OXT3YKPCIVKobVwiMi06YWnjrz5Bsu0kisl88AM6DAI1IVnrz+evl05+XaRRpC0K7+MXVpiVSDurTqLFBN8SI1YYJ1HKmFR6QqPHne8HLg8XLtIg3BevK0NXMJxR2RwFJoEAkwG7AWnpC6tESqwZPHkQKPSIAFrYUnb3TOE6kCBR4R8ZagtfDkjTdP1CJSOS8HHp24RGRa8gELeCI14snrrwKPSICFi8HaWyrv5TOeSOPw5PXXy4d/zu0CRLwuXPDmiWumCh6dTivSYDx53vBy4Mm7XYCI1wWthaegLnmI0w0AACAASURBVC2Raii4XcBMeDnwqIVHpEKRgLXw5Lx8xhNpHFm3C5gJLx/+CjwiFQp78j5t5lIhtfCIVIECT50p8IhUKBKwjuGRMGG3axDxAU9ef70ceAJ2qhapviANWs5DKRvy9DlPpFGohafOPJkwRRpJNB+cwDMepuh2DSI+ULC91pPHkgKPSIAlUsEZ0zIeCtaMNJEa8ey118uBJ+V2ASJelxwLTuAZDivwiFSBJ7uzwNuBZ9TtAkS8Ljnu6XPAtIxEgtN9J1JD424XMFNePtmNuV2AiNc1pYIza2kkrMAjUgWebWxQ4BEJsHiaMKVgBIH+aHC670RqSIGn7qzNoqnpIhUxYGI5by4TP117Ygo8IlXg2cYG7wYeh2ffeJFGEcsGY7r2vjgRt2sQ8QG18LhEgUekQom0/2cvFcH2RxV4RKpAgcclnn3jRRpF65DbFdTeSJhCUTuli1SqaHutZ5eE8XrgGXa7ABGvax90u4LaOxINxjglkRrzdCOD1wPPMbcLEPG69gH/d/UciAZjJppIjQ24XUAlvB54PP3mizSC5lGifp+avi/m759PpE483cjg7cBjbQZtMSFSkZDFNKX8vcTDs03BWWBRpIYUeFzm6f8AkUbQMuLvMS5bE0TdrkHEBzzdq6LAIyK0Dfm3y+dYmFwqrBYekQplba/19FIwfgg8nk6cIo2gfcAX54IT2pnwd+uVSJ14vnHBDyc5z/8niLit64h/u3yeTfi39UqkjjzfuOCHwDOE9tQSqUg8QySe9udxtEkDlkWq4YjbBVTK+4HHWgsccrsMEa/rOObPrp+tTcTcrkHEBw66XUClvB94HAfcLkDE67p9eNtwNEIuE/LNeU7ELWNeH7AM/gk8/W4XIOJ13f3+W3F5U5M/W61E6swXjQp+CTyHQSc2kUq0jhAN5ym6XUc1Pd7sdgUivuD57izAJ3d01pYw5jAw3+1SRLysfYDcwBya6vmahVKklC/GS7liopQrxkslG8GYEgCGEsZYABNyPmcMFoy1BovBmpApEg1nQ/FwJmyMfcGO6OubNX5HpAoUeBrMARR4RCrSdZjSwJzqPFe+GCuO5joLqXxLKVNosel8C+lCM9li0mQLTeFCKRYqlKIhSyhEVVqbrQ2bfDEWzhaj4WwplSA/1L4gTPRQjughQ7Q/RPRQmMiQn857IrU2bnvtiNtFVIOfDnyN4xGp0Nx9RLetm96/sdbYsVx7YSjbkx/OdNvhbHdoLNcRzRWbIlDPKeHGFG0snC7EwulCK1vji4oMn5d88bdli8QOFIjvKhLfDonnQiS2xwhl/NLFL1JNvmjdAX8FnkM46/H4dgE1kVprHyIWy5DPnWTvqVwxXhzOdOWHMrNLw9luRrKzwql8a9QSjtJgx96ulrknDjA2Hia7LEx2GfDS5z8fHsoT21ckvqNI4jlD81MxIgN+OkeKzMQ+twuoFv8czNYWMWYfsMztUkS8rOsQ+YNLnPCSLSSKh1OLsofHF3EsNS+aLTZHqWurzcyUMKX9ye74tP5RsSNKuiNK+uznPxc5mqPpmQLNjxmaH48RGWn4n12kiiywx+0iqsU/gcexCwUekRkrEi2FdqzOP5VoTx1NzQ+n8u1x4MXdQg3ucKIjWwhFKh98XeiOMXpNjNFrnL9HDuVIPl2geb2heX2McEoBSPys3/bajNtFVIvfAs8eoIR/ptuL1Fya7sJBXpLr57LQIKti+cOmde+827Ah69njaHtrjeYvFObEGJkTY+Q6oGSJ7c/S+qsC7XfGiB5pqC49kSrY7XYB1eSvwGNtFmP60WwtkVPK0lbcx8uy+3hZeJRlcSadCyIlmDU+K32s9Vhdp6dXiwW7vXVBHaajhwy5RXGOvT3OsbdDbE+W1vsKtN8VJ3rUX+dWCapdbhdQTX48KHehwCPyIgXipQNcndnLy80gqxIQPmlX1YKBBfZY67F6llc1x+Jt2Uwknqj7C+cWxzl2Y5xjN0JsT4a2e0u03R1T+BGPGvDLdPQJfjwQdwFXuF2ESCOwGNvPZZk9vIqjnBO3RKc0Hmfe4Lz4hsUbLAZz+u9uLDta5pfcroHc4gRH3wlH3wmxXRk6flai/c4EobxnuwklcHzVnQV+DDzWjmHMUaDb7VJE3JKlrbiTN2T28KpYjvZpd03FirFwR6ojM9Q8VP+Wkgpta1vYWKsr55YmOPwhOPLuIu33jDPr+xrvI16wy+0Cqs1/gcexHQUeCaAhVmaf43eKh7g4YYlUtJPUvMF5paHmoWqVVheDsZbseLRpetPR68Umwwy9ppmh6y1Nm9PM+gG0POrJcVLie8O21x5xu4hq82vg2QZcjGZrSQCUCNv9XJvewRtDoyytWovMomOL4lsWbClZ453ZWlvbFnlg89OQIb22if1rIXwsR8fP83T+OKEp7tJAtrldQC34M/BYmyovQrjY7VJEaqVItLSdN6V38oZ4ntaqr5UTL8TDPcM9qcMdhz2xDk8JU9rSvqQxW3dOptgV49iNMY7dUKL9l+N0fyOhxQ2lASjweMxWFHjEh4pESzt4Y2Y7b4oVaKmo2+p0zjh0Ruhwx+FavkTV7Ev2ZHLhqQ3KbjzREMOvbGb4pSU67hin69sKPuKWfttrR90uohb8HHh2AxnAc4MuRU6kRNju4A3p7bwlVosWnRPpGutKNGWbcul4uqKBwAdHD3b+ctcvb84X820AC9oW3Pfy5S+/+9ljzy58ZP8jN5ZsKWowxYvmX/Sd1T2rd83kNTZ1LPXcjLIXi4YYek0zwy8v0X57iu7vxAmPKfhIPW1xu4Ba8W/gsbaEMduAae79LNJYSoTtTl6f3s6bozna696CsfTI0sLmhZsrCjzhULh04bwLv7+qe9We0exo/Aebf3DLjsEdm9cfWP/ms7rP+vFF8y965rEDj539+MHH37y6Z/Xnp/v86XAsv7d5tn9ubmwsxNDrkgy/skjHz8fp+q7G+Eg95IEdbhdRK/4NPI6tKPCIh+3jpalNvCeao8O1rpqlR5Ymts7fWiqFSjMevDy7efbw7ObZwwCt8dZsU6Tp4FBmqAMgV8w1AWQL2aZ4JD6jaWHPtS7IWWP8N9XbxsMMvrGZoeuLdP5knK7vNGktH6mh7bbXFtwuolb8HXisHdCaPOJFIyzJPclHSyOscH1MSqQUCc0dmps6MOtAVWrZP7K/K5VPLVo5a+XOjkTHv9+7694/2nJ0y1sA86oVr/rb6T6fBft0x7LGWnun2mwizMBbmhm+Ls+cr2RpfVDT2aUWNrldQC0F4U7habcLEJmqPMnik3wkdR//EB1hRcN00ZzRf0ZVbo7GcmPxu3fe/cHVPau/1xpvzWw8tPGa1T2rv/ee89/zqbO6z/revbvvfdd0n/NAU1dmNNbsv9adEyl2RjnwZ03s+WyG3Jy82+WIrxy0vfao20XUUhACz3NAyu0iRE5nJ69N3cX/s/u4LgnhhhqA255uj7WkWzKVPEe+mA//eOuPPzi3Ze7Dly287AmAI6kjl1+64NInAC5fePn6VD61dLrP+3jXmQ31XtVF+uwEO78S5si7xrFh63Y54gsb3C6g1vwfeKwtoVYeaWBDrMzew5ezz/DBZIHmhu1mXn54+Yz3qLLW8qOtP7qpOdZ88BUrXnHnxOejoejwhkMbzgR4sv/Js+Lh+LTmwA9Fm7MHk90N0xJWX9EQA29pZvv/zTN2cdrtasTThoE9bhdRa8baANwcGBMD3gEEo9lbPKFE2G7iPaldvLbhWnROpGiKpdvPvd0Ww8VpzxbaeGjjGQ/ue/DjiUhiP2AB1vSs+c94OJ55sv/JGyw2FDKh/EXzL/rOWd1nTfnEe++cc1Nb2xe7Ps6pISSfSDHvCzEiAw0bmqVhPWB77TNuF1FrwQg8AMZcAZztdhkiAMMsy63nUzbFfE+tDPzUkqfG93Tvqelih1OVCUUL31rxynDJhBo+LNaNyRSZ+8UcbfdqULNMVRb4tp9nZ03wf5fW8zZSvrMUcYvF2C3cmLqfWyNeCzsAKw+ujGEb4zja3LEkq7BzHJsIc/BPmtj/ZylKiRl3QUqgbA5C2IEgBR5rR4GdbpchwTXGgty9fDH7HDckLRFPHnvJXDI6f3C+6+NF8iZcfKrzjICO3ZmCscuT7PjnIqnVWbdLkYZWBHzflTXBkyfdCjzpdgESTNv57dS9/FNkjMWev0iv3r866nYrz6aOpZlcOKqVh0+l2Bll72djHL55HGsaolVOGs5m22vH3S6iXoIVeKw9CuxyuwwJjgLx0sP0pjbznqQl6ovjze1WnoIJFZ+YtdLzwbE+wobBNzWz6x+zWrdHjlMAnnC7iHryxQl4mh5zuwAJBqcL60uFI1zku1lEa/atiRlrXBkj8kzHMrXuTFduaYJdXwoxfJ3WJJMJT9te63r3dD0FL/BYOwBsd7sM8bf9XJW+j38Ip5njyy0PmvJNkQUDCypaiHAmciZcfLzrTM1AmgkbD9P/kST9H1YXl+SAp9wuot6CF3gcjwGawSBVZzF2I783/gSfaCoR93UrxOp9q2OhUqiux9HGzuXZfMibA74bxvD1zez5uwzFlqLbpYhrNtpeG7gB7cE8cVg7DGxxuwzxlyxtxfv5fHY3r2mIdWpqLVFIRJYeWVq3Vp50OJZ/apZmZlVF5swmdn65RHZJzu1SpO6yOMu0BE4wA49jPaBBfFIV5fE6pWGCNZh21YFViWghWpc1PB7pXl0oqHWneoqdUXZ/PqxtKQLnSdtrAxl0g3vysDZNADZLk9obYHX2V9waztERuK1LIqVI6MyDZ9b8xmEg1prZ2r5YY3eqzcbD7L8lwcBvazBzMAwR0NYdCHLgcTwFjLpdhHjXAV6SfpDPRIskfT1e51SWHV6WaMo21fSO8b4552pF5ZoJGY68J8nBj6Q0mNn3fm17bWDHrwY78FhbAH7tdhniTTt4Q+pxPpHwy/o6M2UwZu2+tTUbALureW76cFOn57bh8JyR65Ls+4s0NqzQ40+7bK/d53YRbgr0iRoAa3cDu90uQ7zlad43von3JkF7OQHMG5rX1DXSVfWxIAUTKt0/Z13gugpdk7ogyd7PZChFA9sK4FNF4EG3i3CbAo/jAZxVJ0VOqUTYPsafpXbx+kDMxJqOC3ZeEA0Xw1Vt6dnQuSKdiiQi1XxOOY302ib2/k2WUlyhxz+etL028MM3FHgArB0DHne7DGlsJcL2EXrT/Vzuu5WTqyFRSETO3nt21db2GI4ms+u7Vum9dkPmzCb2fC6nHdd9YRTtIwko8Ey2AWcEu8iLlAjbh+nLHOV8XYBPYfGxxclZo7Mq7toqgb1r3oVYY9Rl6Jbs8gS7b81TTGqBQm97wPZa/R+iwPM8a0vA/W6XIY1nIuwc41xNi56CC3deGA2VQhWdYDd1LE0dTXRooLLbcovi7P77AoU2XTC9aavttXvcLqJRKPBMZu0BYJPbZUjjsBj7CL1phZ2pS+QTkbV71864a2ssksg93L1G73ejyM+Ps/vvixQ6NM7RW8bQQOUXUOB5sYeAYbeLEPdZjH2UW9Lqxpq+pUeXJjvHOqfdtWXB3j33glIxFNa5qZEUZsfY87+L6t7ylPuCuqLyyeikcjxnbZ670eaigbeeT6YPc4nCzgxdsPOCyHS7tra1LUz3J7sCtT2HZ+Tnxtn713mt0+MJm4O+5s6JKPCciLVH0KytQNvI74338xKFnQokc8noWfvPmnLX1kg0mf3V7HMUdhpZdkWCfb3ae6uxjeL0VMhxFHhO7gngsNtFSP3t4A2poOx4XmsrDq9Ito+3n3ZH9SKmdNv8S4y6sjwgdb6zDYU0Igvca3utNsY+AZ1cTsZai9O1pYF6AdLPpelN3KwBs1V04Y4Lw6FS6JRdxA/3rMkMxVtj9apJKjRyXZIjNyr0NJ4nbK894HYRjUqB51SsHUF7bQXGMMtyj/OJOIS19ksVNeeao+fsPuekrTz7kt2ppzuXq/vQawZuSDL4aoWexrEfWO92EY1Mged0rN0CPOt2GVJbaboLD/GZUImYjokaWDSwKLn46OLx4z+fDsfyd867SOvteNXhDzQxeqnG9LhvHLjL9loNKD8Fndyn5lfAUbeLkNookCg9yGeLeVq1Z1MNrduzLtmWavtNS08J7C/mX1zKhaNhN+uSSoQNBz4ZJ7NM05/dU8IJO6cdKxd0CjxTYW0R+AWgXyifsRj7MH3ZFHPVylBjIRsyl267NBItRAsAD/WszRxqmqX33fOiIfb1on23XPOI7bX9bhfhBQo8U+VsMHonzih48YlneG9qEK3qWy+JQiJy8faLCzua54893blc77tfFLti7L+lahvHypTtsr12g9tFeIUCz3Q4W0887HYZUh2HuDi9i9dr+nmdjY4t5e5D73C7DKm21LlNHP3dF43TkpoZBO5xuwgv0ZiF6bJ2A8Z0A2e4XYrMXJruwuP8iaZB19kwsfyfc1mktDWeoHMsxbJ9mp01eqCJ+z59E5nB+QCsu/EbzDmnn3v/4v3kxruINR/j2r5/pm1h48+IOnZDkqZNGZqf0gKStZUCfq6tI6bHWA3qnj5jIsBvAbPdLkWmr0TY/oovZEdZqpNyHeUxpY9xZWEn7eWgaS2vfCBD11Cwu7Z+/kfvZtby57j0j+4nNx4mOxzjkS++hmhynKtvuY37Pn09+XSS6z7zQ7dLnZLQWIFlH4bIgG6oa6MA/Nj22iNuF+I16tKaCWe/rdvQJqOe9DQfTLkddnLkzI3ceMv7ef/vAzzBE103cdOnbuCGT3+YD79vnHFfzVwqgv0852efDzsAxvDLS2OkEsG9Sx092MTYwTO55A/uByDWXKR1fprBHeey5necna7X/M6DDG4/z80yp6XUEmFvX0F7btWEBe5W2JkZBZ6ZsjYD/AynaVE84iCXp/dwvevjdm7l1us66Dg48fev8/U3X8mVd/47/35LgkTqq3z1Sjfrq7avcnb6Aea/uCUnHw1zxxWGTCyYS+Ef3dxFJDHK7R99Nz+88RZu/+g7SR2LUcy20b3KuaHqWjlMMdvmcqXTk1ua4NCHdG6svodsr93ldhFepcBTCWtHgZ8Dwb1D9ZAUPfkn+WPXx+08wzMd29i27lquvR+gRImDHFz1Lt71OMDLeNmDm9jknTv60/gOK8d/ztKTj9VJJaPceYUlFwneNi6lQpjM0GJWXH8vb/rWpwlFczz8D9e/4HtMCLw4O3T4lc2MXaxFCavnGdtrN7pdhJcp8FTK2mM4a/RoDYoG9xj/s1gk4XpX0Zf40g1v420/CBGyAHvY0xIlmo4TLwEsYclgilSHu1VWx89YMv5dVp2+RW20JcbdlxUphIp1KKtxdCwbJJIYZOWrdwKw6CXrGetfTDg+wtGt7QAc3dpOOD7qZpkz1v9HEa3PUxU70DZHFVPgqQZnuvov3S5DTm4Hb0iNsML1Qcr/xr+tS5IcfQWv2ON2LbX2AHNTX2Hd1LsPBzvi3HNpnqIJzgWya+UI0eQg+x6aA0D/E6tJdh2kc9lTbPqPywHY9B+X07n8KTfLnLFie5T+P9SCrZXZhTNux3utfA1Go+irxdrtGBMHfDX2wg9S9OS38M6GWNF3K1vP2M3uc9/KW88uUYrmySc+z+dvyJNvypINxYmXdrO7M0lyyO1aK7GRWem/5cLpz7460pXg/ovSXPVoghDB2MT13Ju+y6Nf+h888k8RYq1HuPqWb2BLhnv73s/33/YSYs0DXPMXX3W7zBkbvSrJ2F0ZWta7fsPhQXuBO22vDc5NQA1pWnq1GbMGhZ6Gcj+fTw9xZsNNff4hPzzzNm575T/zz1/8EB96/3mc9/j7ef9jH+Nj75jP/H0f42P3ul3jTGyjPfNJrojlCc+8BXnJ/jSXP5HABCT0+F14KM/y94UJZdSrMHUHcNbaCVY3bw3pl6/arN0E3Od2GeLYyWtTjRh2jvde3vvD+7n/FTdww6fTpJs/wAcecLummdhEZ7risAOwe0ETj52tAa9+UeyI0v/76tqaun7gNoWd6lILT60Yswq4GnSH6pYMswp388+mRNz1gcpB8ATdqT4uaSoSqt7v/Bm7Ulz4dFNgurf8buH/ytD8hLq2Tu0w8FPba4O5VEMNqYWnVqzdirPPiRKlSx7nE3mFnfp4hNmpXi6tbtgBeG5pkl9fkKEUoIHMfnbwo2FKcf1fntwBFHZqRoGnlqzdhjN7S6GnzvZzTXqAtQ3fleUHDzA39WkubrKY2rTC7J3fxD2X5AI3Zd2Pip1RDr9fXZUnthtnzI7CTo0o8NSatc8Bd+DsfyJ1UCRaeob3agZiHdzFwtTfcFGyZmFnwqGeBHdeUQjk4oR+M3xdktw8XdRf6DngDo3ZqS0FnnqwdhfwE0B3NnWwjRsyOTqibtfhdz9k+fgXOK9+u50PdsS5/Sob2G0ofCNs6P99BdfnbbC99m5NPa89BZ56sfYw8N+Ap9dXaXRZOgo7eKMGRdZQAVP6R85J/Qtr6r8n2VhzlNuuNow1aTsXL0uf08T4uZq15eyN9ZDbRQSFAk89WTuCE3r63S7FrzbywVyJuH6va2ScSOEWLsvdweL6tewcL52IcPtVYYZadcH0skMfCvLMuwLOgoIb3C4kSHRhqDdrs8BPcfpspYqGWJHt53INVK6RwyRyf8RV9hm63G9By8XC3HZVnF3ztSO3V+Xnxxm6Pojd/GPAf9teu8PtQoJGgccN1hax9m7gCbdL8ZMN/IGl2tOiBYBn6Uj/AdeED9HcOGOjbMjw4AVJHlub0rR1jzpyUzRg09QPAj+0vfaY24UEkQKPm6x9FLgd0HiECu3nqnQjbA7qR/cyP/UJrkikiDbmmkbbliW584q8BjN7UKk1wtG3B6WVZxPOGjvqinWJVlpuBMa0AS8Hut0uxavu4uu5NHNibtfhJzlCpS+zLnMXi9wbrzMd8WyRax/OM2tEwddT8iVWvLdEZMCvS0mUgAdsr93sdiFBpxaeRvD8YOYtbpfiRfu5Kq2wU139JLN/yNUFz4QdgGw8zC+ujLNj4bjbpch0REMcucmvrdxjwI8VdhqDWngajTFn4uy27te7nar7Jf8nO86CuNt1+MW9zE/9A+cmKt4A1E3L96S4aGOCsPXuzxAo+RIr3m2JjDRmt+nM7ADus73Wr2HOcxR4GpExs3C6uDrcLqXR9XNp+jFu0cysKsgSKn6Jc7K/ZKF3WnVOpWU8z1WPFekYVReXF3T8aJw5X6v/2k7VV8DpwtrqdiHyQgo8jcqYCHAJcLbbpTSye/hSZozFuqBVaD/N2T4uCR1spFlYVWEt655Nsea5JkJq7WloJlPkjBsNoayX/5+OAnfZXjvsdiHyYl7+xfI3awtY+2ucNXvG3C6nER3hfIWdChUwpX/njNTvcW3Mf2EHwBg2rmrm9isLjCazblcjp2ATYQbe6OUZWxuA/1LYaVxq4fECY2LAZcBZbpfSSO7jCxlNRZ+5XbRmPscFob20BmPAtylZzt+UYuWuJCG0XlMjCo0WOOOdYUzRS/8/gzhjdQ65XYicmgKPlxizELgaaHG7FLcNsDr7az6ngcozkMeUvsuZme9zRlPNdzlvRF2DWa5cb0hmghH0vGb2V9J0/swL4/KKOIvHPqmNP71BgcdrjIkCFwFrCXCX5MP0po5wkT8G19bRdtoyn+XCxlox2Q3hQonztqQ5Y7fG9jSayNEcK25u9DB6EPiV7bXaDNpDFHi8yphO4Apggdul1FuWtuIdfDME4eC1TszQGJHCNzkr93OWKiRO1jye59KnCswZ8EKLQnDM+1yatl814v9JDmeHc62Z5kEKPF5nzDLgcgLUzbWFd44/x1v9MH215gqY0m0sSX+Ds5oyRNSScTLzDme4aGOIlnSjtywEQ3x7hqUfaaTxeRZnYdjHbK/18sDqQFPg8QNnCvu55YfvFyz8Bf+Wz9ER7C6ZKXic7vQXOSdyhKTeqymxltXb05y9LU6k6KcF8Lxp2XvzxA41wu/uXpxWnUG3C5HKKPD4iTEtOGv3rAB/zkLRQoOnt5/m7JdYZzfS3Uh3yN4Rzxa54JksSw40Yfx5HHmC+wsRHsMJOvtdrEGqSIHHj5zxPRcCy90updp+zWczA5ytC/kJHCOe+y5nFm5nicbpVEPHcI7zNxeYc1TBxw3h4Twr3hnB2Hq/9yngUeBZ26sLpJ8o8PiZMV04wWepy5VURYqe/N18PQIhXXwmOUoi9x3OLNzJomBOM6+1ttEc520uMP+wgk+9LejL0PJYvW5wxoCngC221xbr9JpSRwo8QWBMN85U9sVul1KJp3nf+C5er8HKZYdoyn2bVYV7WKCgUw8t43nO25xnQX+TFi6sk+b1aRb+Ra27sIeBJ4FtWk/H3xR4gsSYHuAcYBkeXMPnDr6Zz9LZCIMYXXWAZPZbrCr9igUay+SGZCrPuVvyLD6g4FNzhRJnvNMSHqvFIPIBnKCzXV1XwaDAE0TGNANrgNWAJ8bDDLEyez+3BnZl5RLYLXRmvs8ZPMocBZ1G0JQusO7ZLIsPJIhqVlfN9Py/FLP+s5rj0vYDz9heu6uKzykeoMATZMaEgZU4O7LPcrmaU9rAh1J7eHXgBuNmCRXvY37mP1jp0809fSBULLF0f4Yzd4bpHA1sKK+Z6P4cyz9Y6fpIWeBZYJM29wwuBR5xGDMfp9VnCdBwd6tBW3tnLy3ZH7O0eDcLE1ktGOgdbaM5Vm/Ps/hgQmv5VNHS388R3z2T0HME2ITTbVWoclXiMQo88kLOzuzLcVp+5rlcDQCDrMo+wN/5/s55lGjhQeZmf8aSyHY6fP/z+lqoWGLZvgxn7grRMeqJbuOGNuv74/R8Y6oTFtLATpzZVkdrWJV4jAKPnJwxrcAZOOGnw60yNvDh8T1c78vZWeNECo8yO/cLFoe0UKBPtY3mWLEnz6KDUZq1Q/uMRPdnWf7BU90Ev5PEPAAAFP5JREFU5HBCznZgvwYhy4ko8MjUODO8luNMbe+s50v7rTsrS6j4OLOzd7DIPMbshKaUB0j7SI7le/Ms7I9q365pWn5zgejRyVvnFIBdOCFnr6aUy+ko8Mj0OS0/i8uP+dRwzM8Aq3O/5nOevzD0k8w+Tk/hIeaGN9AVL2rxRGkez7PkQJ6F/SE6h+Oa4n4a3f+SouuHRWBP+XFACwTKdCjwSGWcjUsX4ISfBUBbNZ/+aT4wvovf8lx31jiRwjPMyj3CHB5mbmyIuO83dZUKRPNF5h7JMe9Iie7BCK1jMQUgoBAqcqwjx8HZlkPtR+2xntvcLkm8S4FHqsuYJmAOMLf8Zw8VLHJ4N1/Nppjf8AN4j5LI76Atv5lOnqInvE2DjqUS4UKJ2QM55h0u0jMQpn00Rtj6f7ZeOl5guDXH0U7LwZ4IRztjPN/lWwS+YS2abSUzosAjteWs9dPD8+Gno/w47ck7S1vxDr7dcFN7x4gUdtFW2EJn8RlmhTcxK5oi2nB1io+YkqV7MMvco0U6RgxtYxGaUxFPh6BMLM9Qa4FjnSWOdoY52hklFzvdcXSbteypS33iO2pml9qytgj0lx8OY0I4XV+d5UdH+c92Jv1OHuLSLODKYoN5TGmARKGfZHE/LaXdtIb20mL20hopd0/p2JH6sSHDka4ER7pe+Pnm8QKdIwU6Rop0jBraxsK0pKKES40RhAqhIplEgfFEifFkidFmw3Cr4WhnjGw8Ckx3MsJiUOCRmVELjzQWYxJAC9D8YX66Os/irk6yppNsqJ2ciVEMxSiZKEUTo2QilEyI0491KIItYWyekE0RLY4StSPESkPE7TAxO0DCDJBggHjoAM3hIyR9MytMAiiRLZDIlGjKlmjKlEimLU1ZSGQNTVlDPBcilgsTLoUwFow1U9oJvmhKFMOWQrhIMWzJRyyFCOQjJXIxGGuyjLaEGG0OMZaMTKHFZrqGrOV7VX5OCQgFHmlYxvB2nPBzSlGKNk7RJiiWDNYWCYVyhCgQCuUJoRlRIlNkSpZQyRKyllCJchiyFMOGfCSEbYhj6ZvWknG7CPEeNctLQzKGVqYQdgDyhE2esBnz4A7wIg3FhgzFkKGxJ3vPxVl/R2RadIGQRjXf7QJEpCHNdbsA8SYFHmlUOqmJyIk0xB5/4j0KPNKoetwuQEQaUpcxGo4h06fAIw2nfDKr635dIuIZIZx1vUSmRYFHGlEXU5hqLiKBpcAj06bAI41I3Vkicipdp/8WkRdS4JFG1O12ASLS0Ga5XYB4jwKPNCK18IjIqbRp4LJMlwKPNJTySazD7TpEpKEZNLFBpkmBRxpNBxqwLCKnp24tmRYFHmk07W4XICKeoMAj06LAI41GgUdEpkKBR6ZFgUcaTZvbBYiIJ2isn0yLAo80GrXwiMhUJI3RNUymTr8s0mgUeERkKgzQ7HYR4h0KPNIwjCEGJNyuQ0Q8o8XtAsQ7FHikkWj8johMhwKPTJkCjzSSpNsFiIintLpdwP/f3r3HWFrfdRx/f9llWZalXFwLAptSql222FoRqJZiqdr0Fks0jakBW5smBlv9QylgjE1qaRoLxGvaqomCjXhpDVWrFUEuLbeyZbntrrDAXgT2yrJs936Zma9/PM+EYXaBc2bmzO95nvN+JSdzOWcOnyGbM5/zuz1qDwuPmsTpLEn9cIRHPbPwqEmOLR1AUqtYeNQzC4+axMIjqR+OCqtnFh41iS9ekvoxr3QAtYeFR03iCI+kflh41DMLj5rEwiOpHxYe9czCoybxxUtSP46KYG7pEGoHC4+axBcuSf3yjZJ6YuFRk/jvUVK/LDzqiX9g1CRzSgeQ1DoWHvXEwqMmsfBI6pd/x9QT/6GoSfz3KEkaCP/AqBEiHN2RJA2OhUdNYeGRJA2MhUeSJHWehUdNMVo6gKRWitIB1A4WHjVCpoVHkjQ4Fh41yVjpAJJaJ0sHUDtYeNQkI6UDSGqdQ6UDqB0sPGoSp7Uk9cvCo55YeNQkFh5J/TpYOoDawcKjJnFKS1K/LDzqiYVHTeILl6R+jGa62UG9sfCoSfaVDiCpVXyTpJ5ZeNQkFh5J/XDBsnpm4VGTWHgk9WN/6QBqDwuPmsTCI6kfu0sHUHtYeNQke0sHkNQqe0oHUHtYeNQkDk9L6ocjPOqZhUdN4giPpH44wqOeWXjUJLvwQoCSeucIj3pm4VFjZDKK79gk9c7Co55ZeNQ0O0oHkNQKY7izU32w8KhpflA6gKRW2JXpFLh6Z+FR01h4JPXixdIB1C4WHjWNhUdSL5z+Vl8sPGoaC4+kXjjCo75YeNQ0u4DR0iEkNd4LpQOoXSw8apR6EeL20jkkNdoYTmmpTxYeNdG20gEkNdqOTMZKh1C7WHjURM+XDiCp0ZzOUt8sPGoiR3gkvZotpQOofSw8aqLtwEjpEJIaa3PpAGofC48ap56bd5RH0pEcxC3pmgILj5rKIWtJR7LFS0poKiw8aiqHrCUdia8NmhILj5pqI7jtVNJhLDyaEguPGimTQ8DW0jkkNcoYHluhKbLwqMk2lA4gqVG2ZrqDU1Nj4VGTWXgkTfRM6QBqLwuPmmwrcKh0CEmNYeHRlFl41Fj1eTwbS+eQ1Ai7M72wsKbOwqOme650AEmN4OiOpsXCo6ZbXzqApEaw8GhaLDxqtEz24KnL0rAbweltTZOFR22wtnQASUVtdDu6psvCozZYVzqApKLWlA6g9rPwqPEy2Y2nLkvDagTX8mkGWHjUFk5rScNpfX2pGWlaLDxqCwuPNJyeLh1A3WDhUSvU01ru0pCGy348i0szxMKjNnmidABJs2pNfeK6NG0WHrXJOuBA6RCSZs1TpQOoOyw8ao1MRvEFUBoW2zPdnamZY+FR2zxeOoCkWbGqdAB1i4VHrZLJi3ipCanrDuJormaYhUdt5CiP1G2rvZSEZpqFR220lmq7qqTuSZzO0gBYeNQ69Tu/laVzSBqIZzPZWTqEusfCo7ZaBQ55Sx3k6I4GwsKjVsrkAB5EKHXNi5k8WzqEusnCozZ7DDyFVeqQh0oHUHdZeNRa9fW11pTOIWlG7MCLBGuALDxqu0dLB5A0Ix7KJEuHUHdZeNRqmWwH1pfOIWladuBorQbMwqMu+D74zlBqsUcc3dGgWXjUevXlJp4snUPSlOzEy0hoFlh41BXLgdHSIST17UFHdzQbLDzqhHrHlgeWSe2yNZOnS4fQcJhbOoA0gx4GzgbmlQ6ifm2bC2+9EkbmwtgcOH853PItWHwlHJhfPWbP8bB4HTzxVfj0BfD191dLt445AF+6CS57ruivoKm4v3QADQ8LjzojkwMRPAqcXzqL+nXSCCz7Y1h8AHbNgSVXwldWwrPXvfSYpZfDux+pPj97G9xzPSzZC587B666DC77ozLZNUVrMtlSOoSGh1Na6poVwK7SIdSvOVRlB2DPHBid8/KXp3XzYd0SuLouPL+9tio7AL+6DnaeNJtpNW0jwAOlQ2i4WHjUKfWV1O8tnUNTsT/g9Z+FxdfDOY/D5eteuu9Lb4eznoA37j/8537/Qjh75ezl1AxYUa+7k2aNhUedk8kzeBhhC81P2HoN/O/VsOZM+KfTXrrvtgvgg8sO/5lrl8Bd74K/u3nWYmq69lKtt5NmlYVHXXUf1bC5WufH9sGPr4ZvnFN9vWIhbDgTfm/Fyx930+nwxV+DG74M5+yZ9ZiaqvvqkVhpVll41En1cPny0jnUq0cXwlPHVp9vPRoeewss3Vx9fd25sPQxWDThj+RdJ8OnfxOu+Vv48NbZz6spWp/pBUJVhru01GUrgDcDLmhtvJUnwO98AvIoyIB3PAhfqEd07jofPnnLyx//mQ/B/uPgmkvhGuCoUdj8xVmPrX4cBO4pHULDKzI94FLdFcGpwIdL55DE3Zk8XjqEhpdTWuq0TDZTjfRIKmejZUelWXg0DJYBL5YOIQ2pEeC7pUNIFh51XiajwJ3AWOks0hB6MJOdpUNIFh4NhUy24a4tabZtyOSx0iEksPBouDwCbC4dQhoS+6lGVqVGsPBoaGSSwF3AocJRpGHwnUz2lg4hjbPwaKjUawk8C0QarJWZ/F/pENJEFh4NnUyeArfISgPyPPC90iGkySw8Glb3AdtKh5A65iDwP5nuiFTzWHg0lOqt6rdRLayUNH0J3J7JrtJBpCOx8Gho1S/Mt1O9UEuanu9l8mzpENIrsfBoqGWyAXigdA6p5VZnegkXNZuFR0OvPhjtydI5pJbaDNxdOoT0Wiw8UuW7wIbSIaSW2Q3c5iJltYGFRwLqF+zbgO2ls0gtMQL8dyb7SgeRemHhkWqZHAT+i+pdq6RXNgbcmskLpYNIvbLwSBNksoeq9BwsnUVqqATuyOS50kGkflh4pEkyeRG4FVyXIB3B3ZmsLR1C6peFRzqCTDZSndFj6ZFesiyTJ0qHkKbCwiO9gkzWAXdg6ZEAHs3kkdIhpKmy8Eivoh66vwNPY9ZwezzTAzrVbhYe6TXUpedOLD0aTqsyPVhQ7Te3dACpDTJ5OgKA9wBRNo00ax51ZEddYeGRelSXngAuxtKj7lueyfLSIaSZEpmO0kv9iOANwM/jGwZ11zIXKKtrLDzSFERwCvB+4JjSWaQZdl8mK0uHkGaahUeaoghOBD4ILCydRZoBo8BdmawpHUQaBAuPNA0RHAd8ADi5dBZpGvZTXQh0S+kg0qBYeKRpimAe8D7gR0pnkaZgB3BLJjtLB5EGycIjzYAIjgIuBJaWziL1YSPVVc+9WK46z8IjzaAIllIVHw/1VNOtproQqJdO0VCw8EgzrN7B9V5gQeks0hGMAfdnsqp0EGk2WXikAYhgAVXpOaV0FmmC3cBtmTxfOog02yw80oC4rkcN8wxwZyYHSgeRSrDwSAMWwVnARXhIocpI4PuenKxhZ+GRZkF9Xs97gNNKZ9FQ2Qvcnsmm0kGk0iw80iypLzz6NuB83MWlwXsKuNct51LFwiPNsggWAT8HnFg6izppH9V28/Wlg0hNYuGRCohgLnAe8FYgCsdRd6yhGtXZXzqI1DQWHqmgerTnIuCHS2dRq+0H7slkbekgUlNZeKTC6rU951Ct7Tm6cBy1S1KdmLzMUR3p1Vl4pIaod3JdCJxZOIraYSvV9JWHCEo9sPBIDRPBmcA7gBMKR1Ez7QceyGR16SBSm1h4pAaqT2leCvwUML9wHDVDAquAB91qLvXPwiM1WATzgJ+g2s01t3AclbOe6rTkF0sHkdrKwiO1QL2+5zzgzbiNfZhspFqQvLV0EKntLDxSi0RwEnAucBYWny7bTDV1tbF0EKkrLDxSC0VwIvCTwI9i8emSTcBDmWwoHUTqGguP1GIRHE91fa4luManrcaAtcBjmWwrHUbqKguP1AERzKc6vHApsKBwHPXmIPA4sDKTPaXDSF1n4ZE6pN7O/gaq4nNG4Tg6sp3ASmB1JodKh5GGhYVH6qgIXkdVfJbgWT6ljVBNWz3pQmSpDAuP1HERzAHeSLXA+QzgqLKJhspmqmtdrXU0RyrLwiMNkQiOoSo/bwJOwx1eg7ADWEc1mvOD0mEkVSw80pCK4Fiq83zeBJxaOE7bPU9VctZnsqN0GEmHs/BIGi8/pwOLqaa9ji2bqPHGqM7MWU9VctxlJTWchUfSYSJYRFV8zqAa/Rn2dT9jVKM4m6gu97A5k5GykST1w8Ij6VVFMBd4/aRb18/6GQVeoCo3m6gKjouOpRaz8EjqWwQLqYrPKcAi4ETaOw12kKrcbJvwcUcmY0VTSZpRFh5JM6LeAXYicFL9cfy2kPJTYqPArvq2c8LH7ZnsLBlM0uyw8EgauPrSFwvq23ETPl8AHE11HbDxjxM/n1yUxurbaH0bAQ4A+yd8nPj5XmCni4olWXgkNVYEkYkvUpKmzcIjSZI6r/S8uiRJ0sBZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCRJUudZeCTNiIi4KCJW9/jYiyPiuUFnkqRxFh5JfYuI9RHxCxO/l5l3Z+aSGXr+GyPiC0f4/kcj4oGI2BMRW+vPPxURMeHnDkbE7ojYFRHLI+LdE37+1yMiI+JPJj3vJfX3b5yJ/JKax8IjqRUi4grgz4DrgFOBU4DLgQuBeRMeem1mLgReB3wVuDki5ky4fw3wKxExd8L3Pg48OcD4kgqz8EiaEZOnqSLi3Ih4uB5p+UZE/PPkUZuIuKIeqdkUEZ+ov/cbwKXAVfVIzbci4gTg88CnMvNfMnNXVh7OzEsz88DkPJmZwD8AJ1OVo3GbgRXA++r/3snAO4F/n8n/H5KaxcIjacZFxDzgm8CNVIXjH4FfmvSwU4ETgNOBTwJfjoiTMvOvgZuoR2oy8xeBnwGOAf6tjwxzgI8B64Atk+7+Wn0fwEfr5z2sNEnqDguPpEH4aWAu8OeZeSgzbwaWTXrMIeDz9f3fBnYDr7QGaBGwLTNHxr8REfdFxI6I2BcRPzvhsZ+JiB318/0p8NnMHJ30fN8ELq5Hjj5GVYAkdZiFR9IgnAZsqKeVxj076TEvTCwwwF5g4Ss83wvAoonrbjLznZl5Yn3fxNey6+vvLwDOA66LiA9MfLLM3Af8J/AHwA9l5r29/2qS2sjCI2kQNgGnj++eqi3u4+dz0tf3U005XdLzE1RWAvcCHzrCQ74GXAH8fR+5JLWUhUfSVB0dEfPHb1RTWOPuB0aB34qIuRFxCXBBH8+9BThr/IvM3AH8IfCViPhIRBwfEUdFxNuB417pSSLibOBdwKoj3P0d4L3AX/SRS1JLWXgkTdW3gX0Tbp8bvyMzDwK/TLUYeQdwGfAf9L4w+G+At9RrdP61fs5rgd8FrqIqRFuAvwKuBu6b8LPju7v2ALcCN9SPe5l6BOj2zNze6y8sqb3i5VPskjQYEfEA8JeZeUPpLJKGjyM8kgYiIt4dEafWU1ofB94G3FI6l6ThNPe1HyJJU7IE+DrVGpu1wEcyc1PZSJKGlVNakiSp85zSkiRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnWfhkSRJnff/ZYschW0ahnsAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fIshhKeCVdxz"
      },
      "source": [
        "### **Save the best model into a file**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "B9hQ5sd0Vh8g"
      },
      "outputs": [],
      "source": [
        "# Define the output file's name\n",
        "output_file = f'RandomForestClassifier_NF_DCM'\n",
        "output_file"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7-6gKurIVpWI"
      },
      "outputs": [],
      "source": [
        "# Export the binary file of the best model\n",
        "#with open(output_file, \"wb\") as f_out:\n",
        " #   pickle.dump(best_model[1], f_out)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4p3zyawQ54lL"
      },
      "source": [
        "## **Create new models with the most important genes**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yl0zewaB54vC"
      },
      "outputs": [],
      "source": [
        "# New datasets with the most important genes\n",
        "most_imp_genes = df_final[\"Feature\"]\n",
        "most_imp_genes.loc[len(most_imp_genes)] = \"etiology\"\n",
        "\n",
        "# Filter dataset with the most important genes\n",
        "df_most_imp_genes = gxData_cpm_labeled[most_imp_genes]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fqY249or_Eu9"
      },
      "outputs": [],
      "source": [
        "# Create the y label dataset\n",
        "y_impgenes = df_most_imp_genes['etiology'].copy()\n",
        "\n",
        "# Encoding the target class label\n",
        "y_impgenes = y_impgenes.map({\"NF\": 0, \"DCM\": 1, \"HCM\": 2, \"PPCM\": 3}).to_numpy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fx-87tOn_Eu_"
      },
      "outputs": [],
      "source": [
        "# Create the feature matrix by deleting the y column and the sample name\n",
        "df_most_imp_genes_def = df_most_imp_genes.drop(['etiology'], axis=1)\n",
        "X_impgenes = df_most_imp_genes_def.to_numpy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 553
        },
        "id": "cM3uDgPH9F0V",
        "outputId": "f10570ef-3d22-4631-b3bd-339276df5660"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-d7b4a018-0340-43b3-982d-9782c94404ce\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>ENSG00000205795</th>\n",
              "      <th>ENSG00000162998</th>\n",
              "      <th>ENSG00000075413</th>\n",
              "      <th>ENSG00000139289</th>\n",
              "      <th>ENSG00000096060</th>\n",
              "      <th>ENSG00000123243</th>\n",
              "      <th>ENSG00000213694</th>\n",
              "      <th>ENSG00000186160</th>\n",
              "      <th>ENSG00000095587</th>\n",
              "      <th>ENSG00000143768</th>\n",
              "      <th>...</th>\n",
              "      <th>ENSG00000079931</th>\n",
              "      <th>ENSG00000177791</th>\n",
              "      <th>ENSG00000065978</th>\n",
              "      <th>ENSG00000115290</th>\n",
              "      <th>ENSG00000197299</th>\n",
              "      <th>ENSG00000196843</th>\n",
              "      <th>ENSG00000197122</th>\n",
              "      <th>ENSG00000196972</th>\n",
              "      <th>ENSG00000198542</th>\n",
              "      <th>etiology</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>7.37</td>\n",
              "      <td>9.97</td>\n",
              "      <td>14.66</td>\n",
              "      <td>12.15</td>\n",
              "      <td>16.50</td>\n",
              "      <td>12.55</td>\n",
              "      <td>13.68</td>\n",
              "      <td>8.05</td>\n",
              "      <td>10.65</td>\n",
              "      <td>4.62</td>\n",
              "      <td>...</td>\n",
              "      <td>9.60</td>\n",
              "      <td>7.97</td>\n",
              "      <td>16.48</td>\n",
              "      <td>11.85</td>\n",
              "      <td>11.73</td>\n",
              "      <td>10.69</td>\n",
              "      <td>10.06</td>\n",
              "      <td>9.15</td>\n",
              "      <td>8.87</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>9.83</td>\n",
              "      <td>11.05</td>\n",
              "      <td>14.27</td>\n",
              "      <td>13.83</td>\n",
              "      <td>14.67</td>\n",
              "      <td>13.61</td>\n",
              "      <td>13.14</td>\n",
              "      <td>5.57</td>\n",
              "      <td>12.76</td>\n",
              "      <td>7.99</td>\n",
              "      <td>...</td>\n",
              "      <td>11.44</td>\n",
              "      <td>7.99</td>\n",
              "      <td>16.71</td>\n",
              "      <td>11.85</td>\n",
              "      <td>11.27</td>\n",
              "      <td>11.24</td>\n",
              "      <td>9.62</td>\n",
              "      <td>9.38</td>\n",
              "      <td>9.44</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>7.07</td>\n",
              "      <td>9.66</td>\n",
              "      <td>14.09</td>\n",
              "      <td>14.90</td>\n",
              "      <td>16.61</td>\n",
              "      <td>12.27</td>\n",
              "      <td>14.16</td>\n",
              "      <td>5.00</td>\n",
              "      <td>12.22</td>\n",
              "      <td>2.67</td>\n",
              "      <td>...</td>\n",
              "      <td>10.81</td>\n",
              "      <td>7.72</td>\n",
              "      <td>17.10</td>\n",
              "      <td>10.28</td>\n",
              "      <td>11.61</td>\n",
              "      <td>13.60</td>\n",
              "      <td>10.40</td>\n",
              "      <td>7.07</td>\n",
              "      <td>6.76</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>8.53</td>\n",
              "      <td>10.54</td>\n",
              "      <td>15.08</td>\n",
              "      <td>12.54</td>\n",
              "      <td>16.96</td>\n",
              "      <td>12.66</td>\n",
              "      <td>13.85</td>\n",
              "      <td>10.00</td>\n",
              "      <td>10.46</td>\n",
              "      <td>5.87</td>\n",
              "      <td>...</td>\n",
              "      <td>10.54</td>\n",
              "      <td>8.41</td>\n",
              "      <td>16.73</td>\n",
              "      <td>12.49</td>\n",
              "      <td>12.28</td>\n",
              "      <td>10.58</td>\n",
              "      <td>10.38</td>\n",
              "      <td>9.68</td>\n",
              "      <td>10.31</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>8.24</td>\n",
              "      <td>10.90</td>\n",
              "      <td>14.93</td>\n",
              "      <td>12.66</td>\n",
              "      <td>18.18</td>\n",
              "      <td>12.69</td>\n",
              "      <td>13.73</td>\n",
              "      <td>6.93</td>\n",
              "      <td>10.85</td>\n",
              "      <td>8.29</td>\n",
              "      <td>...</td>\n",
              "      <td>10.91</td>\n",
              "      <td>9.01</td>\n",
              "      <td>16.92</td>\n",
              "      <td>12.36</td>\n",
              "      <td>12.19</td>\n",
              "      <td>11.49</td>\n",
              "      <td>9.89</td>\n",
              "      <td>9.79</td>\n",
              "      <td>10.54</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>361</th>\n",
              "      <td>11.00</td>\n",
              "      <td>13.02</td>\n",
              "      <td>14.01</td>\n",
              "      <td>13.35</td>\n",
              "      <td>13.48</td>\n",
              "      <td>13.67</td>\n",
              "      <td>10.73</td>\n",
              "      <td>6.59</td>\n",
              "      <td>12.43</td>\n",
              "      <td>9.66</td>\n",
              "      <td>...</td>\n",
              "      <td>10.73</td>\n",
              "      <td>10.90</td>\n",
              "      <td>16.66</td>\n",
              "      <td>9.85</td>\n",
              "      <td>8.84</td>\n",
              "      <td>14.40</td>\n",
              "      <td>9.40</td>\n",
              "      <td>9.92</td>\n",
              "      <td>10.97</td>\n",
              "      <td>HCM</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>362</th>\n",
              "      <td>10.39</td>\n",
              "      <td>11.79</td>\n",
              "      <td>13.98</td>\n",
              "      <td>14.16</td>\n",
              "      <td>14.19</td>\n",
              "      <td>13.63</td>\n",
              "      <td>12.78</td>\n",
              "      <td>4.81</td>\n",
              "      <td>12.07</td>\n",
              "      <td>8.85</td>\n",
              "      <td>...</td>\n",
              "      <td>11.32</td>\n",
              "      <td>12.04</td>\n",
              "      <td>16.66</td>\n",
              "      <td>10.77</td>\n",
              "      <td>9.83</td>\n",
              "      <td>12.34</td>\n",
              "      <td>10.25</td>\n",
              "      <td>9.84</td>\n",
              "      <td>12.04</td>\n",
              "      <td>DCM</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>363</th>\n",
              "      <td>8.29</td>\n",
              "      <td>10.83</td>\n",
              "      <td>15.29</td>\n",
              "      <td>12.62</td>\n",
              "      <td>16.90</td>\n",
              "      <td>12.04</td>\n",
              "      <td>13.79</td>\n",
              "      <td>9.39</td>\n",
              "      <td>10.29</td>\n",
              "      <td>6.27</td>\n",
              "      <td>...</td>\n",
              "      <td>9.20</td>\n",
              "      <td>8.29</td>\n",
              "      <td>16.34</td>\n",
              "      <td>12.72</td>\n",
              "      <td>12.72</td>\n",
              "      <td>10.15</td>\n",
              "      <td>9.60</td>\n",
              "      <td>9.28</td>\n",
              "      <td>10.09</td>\n",
              "      <td>NF</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>364</th>\n",
              "      <td>10.37</td>\n",
              "      <td>13.50</td>\n",
              "      <td>13.93</td>\n",
              "      <td>14.41</td>\n",
              "      <td>12.98</td>\n",
              "      <td>13.79</td>\n",
              "      <td>11.70</td>\n",
              "      <td>3.66</td>\n",
              "      <td>12.62</td>\n",
              "      <td>9.10</td>\n",
              "      <td>...</td>\n",
              "      <td>12.26</td>\n",
              "      <td>9.45</td>\n",
              "      <td>16.70</td>\n",
              "      <td>11.85</td>\n",
              "      <td>9.38</td>\n",
              "      <td>12.89</td>\n",
              "      <td>10.19</td>\n",
              "      <td>10.01</td>\n",
              "      <td>12.12</td>\n",
              "      <td>DCM</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>365</th>\n",
              "      <td>10.76</td>\n",
              "      <td>11.77</td>\n",
              "      <td>14.23</td>\n",
              "      <td>14.31</td>\n",
              "      <td>13.09</td>\n",
              "      <td>13.48</td>\n",
              "      <td>11.48</td>\n",
              "      <td>5.84</td>\n",
              "      <td>12.55</td>\n",
              "      <td>9.93</td>\n",
              "      <td>...</td>\n",
              "      <td>10.57</td>\n",
              "      <td>10.83</td>\n",
              "      <td>16.43</td>\n",
              "      <td>11.35</td>\n",
              "      <td>8.53</td>\n",
              "      <td>14.02</td>\n",
              "      <td>9.62</td>\n",
              "      <td>10.13</td>\n",
              "      <td>10.60</td>\n",
              "      <td>DCM</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>366 rows × 281 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d7b4a018-0340-43b3-982d-9782c94404ce')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-d7b4a018-0340-43b3-982d-9782c94404ce button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-d7b4a018-0340-43b3-982d-9782c94404ce');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "     ENSG00000205795  ENSG00000162998  ENSG00000075413  ENSG00000139289  \\\n",
              "0               7.37             9.97            14.66            12.15   \n",
              "1               9.83            11.05            14.27            13.83   \n",
              "2               7.07             9.66            14.09            14.90   \n",
              "3               8.53            10.54            15.08            12.54   \n",
              "4               8.24            10.90            14.93            12.66   \n",
              "..               ...              ...              ...              ...   \n",
              "361            11.00            13.02            14.01            13.35   \n",
              "362            10.39            11.79            13.98            14.16   \n",
              "363             8.29            10.83            15.29            12.62   \n",
              "364            10.37            13.50            13.93            14.41   \n",
              "365            10.76            11.77            14.23            14.31   \n",
              "\n",
              "     ENSG00000096060  ENSG00000123243  ENSG00000213694  ENSG00000186160  \\\n",
              "0              16.50            12.55            13.68             8.05   \n",
              "1              14.67            13.61            13.14             5.57   \n",
              "2              16.61            12.27            14.16             5.00   \n",
              "3              16.96            12.66            13.85            10.00   \n",
              "4              18.18            12.69            13.73             6.93   \n",
              "..               ...              ...              ...              ...   \n",
              "361            13.48            13.67            10.73             6.59   \n",
              "362            14.19            13.63            12.78             4.81   \n",
              "363            16.90            12.04            13.79             9.39   \n",
              "364            12.98            13.79            11.70             3.66   \n",
              "365            13.09            13.48            11.48             5.84   \n",
              "\n",
              "     ENSG00000095587  ENSG00000143768  ...  ENSG00000079931  ENSG00000177791  \\\n",
              "0              10.65             4.62  ...             9.60             7.97   \n",
              "1              12.76             7.99  ...            11.44             7.99   \n",
              "2              12.22             2.67  ...            10.81             7.72   \n",
              "3              10.46             5.87  ...            10.54             8.41   \n",
              "4              10.85             8.29  ...            10.91             9.01   \n",
              "..               ...              ...  ...              ...              ...   \n",
              "361            12.43             9.66  ...            10.73            10.90   \n",
              "362            12.07             8.85  ...            11.32            12.04   \n",
              "363            10.29             6.27  ...             9.20             8.29   \n",
              "364            12.62             9.10  ...            12.26             9.45   \n",
              "365            12.55             9.93  ...            10.57            10.83   \n",
              "\n",
              "     ENSG00000065978  ENSG00000115290  ENSG00000197299  ENSG00000196843  \\\n",
              "0              16.48            11.85            11.73            10.69   \n",
              "1              16.71            11.85            11.27            11.24   \n",
              "2              17.10            10.28            11.61            13.60   \n",
              "3              16.73            12.49            12.28            10.58   \n",
              "4              16.92            12.36            12.19            11.49   \n",
              "..               ...              ...              ...              ...   \n",
              "361            16.66             9.85             8.84            14.40   \n",
              "362            16.66            10.77             9.83            12.34   \n",
              "363            16.34            12.72            12.72            10.15   \n",
              "364            16.70            11.85             9.38            12.89   \n",
              "365            16.43            11.35             8.53            14.02   \n",
              "\n",
              "     ENSG00000197122  ENSG00000196972  ENSG00000198542  etiology  \n",
              "0              10.06             9.15             8.87        NF  \n",
              "1               9.62             9.38             9.44        NF  \n",
              "2              10.40             7.07             6.76        NF  \n",
              "3              10.38             9.68            10.31        NF  \n",
              "4               9.89             9.79            10.54        NF  \n",
              "..               ...              ...              ...       ...  \n",
              "361             9.40             9.92            10.97       HCM  \n",
              "362            10.25             9.84            12.04       DCM  \n",
              "363             9.60             9.28            10.09        NF  \n",
              "364            10.19            10.01            12.12       DCM  \n",
              "365             9.62            10.13            10.60       DCM  \n",
              "\n",
              "[366 rows x 281 columns]"
            ]
          },
          "execution_count": 149,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df_most_imp_genes"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZITP-KlBAA36"
      },
      "outputs": [],
      "source": [
        "# Filter external datasets with most important genes\n",
        "ext_gxData1_cpm_ig = ext_gxData1_cpm[most_imp_genes]\n",
        "ext_gxData2_cpm_ig = ext_gxData2_cpm[most_imp_genes]\n",
        "ext_gxData3_cpm_ig = ext_gxData3_cpm[most_imp_genes]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9V9zu9u6A1wl"
      },
      "outputs": [],
      "source": [
        "# Create the y label dataset\n",
        "y_ext1_ig = ext_gxData1_cpm_ig['etiology'].copy()\n",
        "y_ext2_ig = ext_gxData2_cpm_ig['etiology'].copy()\n",
        "y_ext3_ig = ext_gxData3_cpm_ig['etiology'].copy()\n",
        "\n",
        "# Encoding the target class label\n",
        "y_ext1_ig = y_ext1_ig.map({\"DCM\": 1, \"NF\": 0}).to_numpy()\n",
        "y_ext2_ig = y_ext2_ig.map({\"DCM\": 1, \"NF\": 0}).to_numpy()\n",
        "y_ext3_ig = y_ext3_ig.map({\"DCM\": 1, \"NF\": 0}).to_numpy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "--CbdlIlA1wo"
      },
      "outputs": [],
      "source": [
        "# Create the feature matrix by deleting the y column\n",
        "final_df_NFDCM_ext1_ig = ext_gxData1_cpm_ig.drop(['etiology'], axis=1)\n",
        "final_df_NFDCM_ext2_ig = ext_gxData2_cpm_ig.drop(['etiology'], axis=1)\n",
        "final_df_NFDCM_ext3_ig = ext_gxData3_cpm_ig.drop(['etiology'], axis=1)\n",
        "\n",
        "X_ext1_ig = final_df_NFDCM_ext1_ig.to_numpy()\n",
        "X_ext2_ig = final_df_NFDCM_ext2_ig.to_numpy()\n",
        "X_ext3_ig = final_df_NFDCM_ext3_ig.to_numpy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uJfo2FLM_jJh"
      },
      "outputs": [],
      "source": [
        "# Run the ML classifiers with LazyPredict     \n",
        "# Data splitting\n",
        "X_train_ig, X_test_ig, y_train_ig, y_test_ig = train_test_split(X_impgenes, y_impgenes, test_size=0.2, \n",
        "                                                    random_state =42, \n",
        "                                                    stratify=y_impgenes)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dAJ_EWMd_jJj",
        "outputId": "b24844bf-aac8-45ae-e504-4d653a9c1ca5"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 29/29 [00:04<00:00,  6.58it/s]\n"
          ]
        }
      ],
      "source": [
        "# Define and build the Lazyclassifier\n",
        "clf = LazyClassifier(verbose=0,ignore_warnings=True, custom_metric=matthews_corrcoef)\n",
        "\n",
        "models, predictions = clf.fit(X_train_ig, X_test_ig, y_train_ig, y_test_ig)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "xatXm70z_jJl",
        "outputId": "69dbb91c-5492-439e-c138-53ff322a2fbb"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-3d4497ce-4966-4fd0-bc01-2f624a2a2af1\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Balanced Accuracy</th>\n",
              "      <th>ROC AUC</th>\n",
              "      <th>F1 Score</th>\n",
              "      <th>matthews_corrcoef</th>\n",
              "      <th>Time Taken</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Model</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>LinearSVC</th>\n",
              "      <td>0.95</td>\n",
              "      <td>0.66</td>\n",
              "      <td>None</td>\n",
              "      <td>0.94</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>PassiveAggressiveClassifier</th>\n",
              "      <td>0.95</td>\n",
              "      <td>0.66</td>\n",
              "      <td>None</td>\n",
              "      <td>0.94</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Perceptron</th>\n",
              "      <td>0.93</td>\n",
              "      <td>0.62</td>\n",
              "      <td>None</td>\n",
              "      <td>0.92</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RidgeClassifierCV</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.60</td>\n",
              "      <td>None</td>\n",
              "      <td>0.90</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LinearDiscriminantAnalysis</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.60</td>\n",
              "      <td>None</td>\n",
              "      <td>0.90</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RidgeClassifier</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.60</td>\n",
              "      <td>None</td>\n",
              "      <td>0.90</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BernoulliNB</th>\n",
              "      <td>0.86</td>\n",
              "      <td>0.58</td>\n",
              "      <td>None</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.77</td>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>GaussianNB</th>\n",
              "      <td>0.92</td>\n",
              "      <td>0.58</td>\n",
              "      <td>None</td>\n",
              "      <td>0.90</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LogisticRegression</th>\n",
              "      <td>0.92</td>\n",
              "      <td>0.58</td>\n",
              "      <td>None</td>\n",
              "      <td>0.90</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DecisionTreeClassifier</th>\n",
              "      <td>0.82</td>\n",
              "      <td>0.56</td>\n",
              "      <td>None</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.71</td>\n",
              "      <td>0.04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SVC</th>\n",
              "      <td>0.92</td>\n",
              "      <td>0.54</td>\n",
              "      <td>None</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SGDClassifier</th>\n",
              "      <td>0.92</td>\n",
              "      <td>0.54</td>\n",
              "      <td>None</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>KNeighborsClassifier</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.53</td>\n",
              "      <td>None</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ExtraTreeClassifier</th>\n",
              "      <td>0.82</td>\n",
              "      <td>0.52</td>\n",
              "      <td>None</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.71</td>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>NearestCentroid</th>\n",
              "      <td>0.78</td>\n",
              "      <td>0.50</td>\n",
              "      <td>None</td>\n",
              "      <td>0.80</td>\n",
              "      <td>0.65</td>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ExtraTreesClassifier</th>\n",
              "      <td>0.91</td>\n",
              "      <td>0.50</td>\n",
              "      <td>None</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.84</td>\n",
              "      <td>0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>XGBClassifier</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>None</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.82</td>\n",
              "      <td>0.97</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RandomForestClassifier</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>None</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.82</td>\n",
              "      <td>0.26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LGBMClassifier</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>None</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.82</td>\n",
              "      <td>1.46</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>CalibratedClassifierCV</th>\n",
              "      <td>0.89</td>\n",
              "      <td>0.49</td>\n",
              "      <td>None</td>\n",
              "      <td>0.86</td>\n",
              "      <td>0.81</td>\n",
              "      <td>0.35</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BaggingClassifier</th>\n",
              "      <td>0.86</td>\n",
              "      <td>0.48</td>\n",
              "      <td>None</td>\n",
              "      <td>0.83</td>\n",
              "      <td>0.77</td>\n",
              "      <td>0.17</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>AdaBoostClassifier</th>\n",
              "      <td>0.74</td>\n",
              "      <td>0.41</td>\n",
              "      <td>None</td>\n",
              "      <td>0.71</td>\n",
              "      <td>0.57</td>\n",
              "      <td>0.38</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LabelSpreading</th>\n",
              "      <td>0.46</td>\n",
              "      <td>0.25</td>\n",
              "      <td>None</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>LabelPropagation</th>\n",
              "      <td>0.46</td>\n",
              "      <td>0.25</td>\n",
              "      <td>None</td>\n",
              "      <td>0.29</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>DummyClassifier</th>\n",
              "      <td>0.45</td>\n",
              "      <td>0.25</td>\n",
              "      <td>None</td>\n",
              "      <td>0.28</td>\n",
              "      <td>0.00</td>\n",
              "      <td>0.01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>QuadraticDiscriminantAnalysis</th>\n",
              "      <td>0.41</td>\n",
              "      <td>0.22</td>\n",
              "      <td>None</td>\n",
              "      <td>0.41</td>\n",
              "      <td>0.02</td>\n",
              "      <td>0.03</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3d4497ce-4966-4fd0-bc01-2f624a2a2af1')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-3d4497ce-4966-4fd0-bc01-2f624a2a2af1 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-3d4497ce-4966-4fd0-bc01-2f624a2a2af1');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "                               Accuracy  Balanced Accuracy ROC AUC  F1 Score  \\\n",
              "Model                                                                          \n",
              "LinearSVC                          0.95               0.66    None      0.94   \n",
              "PassiveAggressiveClassifier        0.95               0.66    None      0.94   \n",
              "Perceptron                         0.93               0.62    None      0.92   \n",
              "RidgeClassifierCV                  0.91               0.60    None      0.90   \n",
              "LinearDiscriminantAnalysis         0.91               0.60    None      0.90   \n",
              "RidgeClassifier                    0.91               0.60    None      0.90   \n",
              "BernoulliNB                        0.86               0.58    None      0.88   \n",
              "GaussianNB                         0.92               0.58    None      0.90   \n",
              "LogisticRegression                 0.92               0.58    None      0.90   \n",
              "DecisionTreeClassifier             0.82               0.56    None      0.84   \n",
              "SVC                                0.92               0.54    None      0.89   \n",
              "SGDClassifier                      0.92               0.54    None      0.89   \n",
              "KNeighborsClassifier               0.91               0.53    None      0.88   \n",
              "ExtraTreeClassifier                0.82               0.52    None      0.84   \n",
              "NearestCentroid                    0.78               0.50    None      0.80   \n",
              "ExtraTreesClassifier               0.91               0.50    None      0.86   \n",
              "XGBClassifier                      0.89               0.49    None      0.85   \n",
              "RandomForestClassifier             0.89               0.49    None      0.85   \n",
              "LGBMClassifier                     0.89               0.49    None      0.85   \n",
              "CalibratedClassifierCV             0.89               0.49    None      0.86   \n",
              "BaggingClassifier                  0.86               0.48    None      0.83   \n",
              "AdaBoostClassifier                 0.74               0.41    None      0.71   \n",
              "LabelSpreading                     0.46               0.25    None      0.29   \n",
              "LabelPropagation                   0.46               0.25    None      0.29   \n",
              "DummyClassifier                    0.45               0.25    None      0.28   \n",
              "QuadraticDiscriminantAnalysis      0.41               0.22    None      0.41   \n",
              "\n",
              "                               matthews_corrcoef  Time Taken  \n",
              "Model                                                         \n",
              "LinearSVC                                   0.91        0.14  \n",
              "PassiveAggressiveClassifier                 0.91        0.03  \n",
              "Perceptron                                  0.88        0.02  \n",
              "RidgeClassifierCV                           0.84        0.03  \n",
              "LinearDiscriminantAnalysis                  0.84        0.04  \n",
              "RidgeClassifier                             0.84        0.02  \n",
              "BernoulliNB                                 0.77        0.01  \n",
              "GaussianNB                                  0.86        0.01  \n",
              "LogisticRegression                          0.86        0.06  \n",
              "DecisionTreeClassifier                      0.71        0.04  \n",
              "SVC                                         0.86        0.08  \n",
              "SGDClassifier                               0.86        0.03  \n",
              "KNeighborsClassifier                        0.84        0.01  \n",
              "ExtraTreeClassifier                         0.71        0.01  \n",
              "NearestCentroid                             0.65        0.01  \n",
              "ExtraTreesClassifier                        0.84        0.13  \n",
              "XGBClassifier                               0.82        0.97  \n",
              "RandomForestClassifier                      0.82        0.26  \n",
              "LGBMClassifier                              0.82        1.46  \n",
              "CalibratedClassifierCV                      0.81        0.35  \n",
              "BaggingClassifier                           0.77        0.17  \n",
              "AdaBoostClassifier                          0.57        0.38  \n",
              "LabelSpreading                              0.00        0.02  \n",
              "LabelPropagation                            0.00        0.02  \n",
              "DummyClassifier                             0.00        0.01  \n",
              "QuadraticDiscriminantAnalysis               0.02        0.03  "
            ]
          },
          "execution_count": 135,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "models"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gjmmdqa9BcmX",
        "outputId": "d2575edb-8606-446b-84b7-e77b67ed0467"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Pipeline(steps=[('standardscaler', StandardScaler()),\n",
              "                ('xgbclassifier',\n",
              "                 XGBClassifier(objective='multi:softprob', random_state=42))])"
            ]
          },
          "execution_count": 154,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Implementation of the best models\n",
        "best_model1_ig = make_pipeline(StandardScaler(), RandomForestClassifier(random_state=42))\n",
        "best_model1_ig.fit(X_train_ig, y_train_ig)\n",
        "\n",
        "best_model2_ig = make_pipeline(StandardScaler(), LGBMClassifier(random_state=42))\n",
        "best_model2_ig.fit(X_train_ig, y_train_ig)\n",
        "\n",
        "best_model3_ig = make_pipeline(StandardScaler(), XGBClassifier(random_state=42))\n",
        "best_model3_ig.fit(X_train_ig, y_train_ig)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ebvChuQ0T5ix"
      },
      "outputs": [],
      "source": [
        "models_ig = [best_model1_ig, best_model2_ig, best_model3_ig]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HcfKNWssT_-j"
      },
      "outputs": [],
      "source": [
        "# Dataset to store performance metrics\n",
        "results_bestMLmodel_ig = pd.DataFrame()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "K06fZyhPT95M"
      },
      "outputs": [],
      "source": [
        "for i in models_ig:\n",
        "  # Dataset to store performance metrics\n",
        "  results_bestMLmodel_temp = pd.DataFrame()\n",
        "  \n",
        "  # Predicting values with the best model on test and external datasets\n",
        "  y_test_pred_ig = i.predict(X_test_ig)\n",
        "  y_ext_pred1_ig = i.predict(X_ext1_ig)\n",
        "  y_ext_pred2_ig = i.predict(X_ext2_ig)\n",
        "  y_ext_pred3_ig = i.predict(X_ext3_ig)\n",
        "\n",
        "  # Calculate performance metrics of test and external datasets\n",
        "  results_bestMLmodel_temp[\"Performance_metric\"] = ['Accuracy', 'Precision', \n",
        "                                              'Recall', 'F1score', 'MCC']\n",
        "\n",
        "  results_bestMLmodel_temp[\"Model\"] = [type(i[1]).__name__]*len(results_bestMLmodel_temp)\n",
        "\n",
        "  results_bestMLmodel_temp[\"Test_dataset\"] = performance_metrics(y_test_ig, y_test_pred_ig, i, X_test_ig)\n",
        "\n",
        "  results_bestMLmodel_temp[\"External_dataset_gse116260\"] = performance_metrics(y_ext1_ig, y_ext_pred1_ig, i, X_ext1_ig)\n",
        "\n",
        "  results_bestMLmodel_temp[\"External_dataset_gse46224\"] = performance_metrics(y_ext2_ig, y_ext_pred2_ig, i, X_ext2_ig)\n",
        "\n",
        "  results_bestMLmodel_temp[\"External_dataset_gse55296\"] = performance_metrics(y_ext3_ig, y_ext_pred3_ig, i, X_ext3_ig)\n",
        "\n",
        "  # Append results to the general dataset \n",
        "  results_bestMLmodel_ig = pd.concat([results_bestMLmodel_ig, results_bestMLmodel_temp], ignore_index=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 600
        },
        "id": "6x5y7ef8VCbX",
        "outputId": "890484f6-bcc3-4125-e952-4668d606e9ac"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-706e6e6c-c630-4c19-9bcb-624de1e15365\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Performance_metric</th>\n",
              "      <th>Model</th>\n",
              "      <th>Test_dataset</th>\n",
              "      <th>External_dataset_gse116260</th>\n",
              "      <th>External_dataset_gse46224</th>\n",
              "      <th>External_dataset_gse55296</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Accuracy</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.37</td>\n",
              "      <td>0.69</td>\n",
              "      <td>0.43</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Accuracy</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.59</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Accuracy</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.51</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.48</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>F1score</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.43</td>\n",
              "      <td>0.71</td>\n",
              "      <td>0.51</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>F1score</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.88</td>\n",
              "      <td>0.59</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.54</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>F1score</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.85</td>\n",
              "      <td>0.53</td>\n",
              "      <td>0.51</td>\n",
              "      <td>0.55</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>MCC</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.82</td>\n",
              "      <td>-0.14</td>\n",
              "      <td>0.42</td>\n",
              "      <td>0.13</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>MCC</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.84</td>\n",
              "      <td>-0.01</td>\n",
              "      <td>-0.13</td>\n",
              "      <td>0.10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>MCC</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.82</td>\n",
              "      <td>-0.07</td>\n",
              "      <td>0.06</td>\n",
              "      <td>0.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>Precision</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.82</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.75</td>\n",
              "      <td>0.61</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>Precision</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.60</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.57</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>Precision</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.82</td>\n",
              "      <td>0.57</td>\n",
              "      <td>0.53</td>\n",
              "      <td>0.66</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>Recall</td>\n",
              "      <td>RandomForestClassifier</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.37</td>\n",
              "      <td>0.69</td>\n",
              "      <td>0.43</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>Recall</td>\n",
              "      <td>LGBMClassifier</td>\n",
              "      <td>0.91</td>\n",
              "      <td>0.59</td>\n",
              "      <td>0.44</td>\n",
              "      <td>0.52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>Recall</td>\n",
              "      <td>XGBClassifier</td>\n",
              "      <td>0.89</td>\n",
              "      <td>0.51</td>\n",
              "      <td>0.50</td>\n",
              "      <td>0.48</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-706e6e6c-c630-4c19-9bcb-624de1e15365')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-706e6e6c-c630-4c19-9bcb-624de1e15365 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-706e6e6c-c630-4c19-9bcb-624de1e15365');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "   Performance_metric                   Model  Test_dataset  \\\n",
              "0            Accuracy  RandomForestClassifier          0.89   \n",
              "1            Accuracy          LGBMClassifier          0.91   \n",
              "2            Accuracy           XGBClassifier          0.89   \n",
              "3             F1score  RandomForestClassifier          0.85   \n",
              "4             F1score          LGBMClassifier          0.88   \n",
              "5             F1score           XGBClassifier          0.85   \n",
              "6                 MCC  RandomForestClassifier          0.82   \n",
              "7                 MCC          LGBMClassifier          0.84   \n",
              "8                 MCC           XGBClassifier          0.82   \n",
              "9           Precision  RandomForestClassifier          0.82   \n",
              "10          Precision          LGBMClassifier          0.91   \n",
              "11          Precision           XGBClassifier          0.82   \n",
              "12             Recall  RandomForestClassifier          0.89   \n",
              "13             Recall          LGBMClassifier          0.91   \n",
              "14             Recall           XGBClassifier          0.89   \n",
              "\n",
              "    External_dataset_gse116260  External_dataset_gse46224  \\\n",
              "0                         0.37                       0.69   \n",
              "1                         0.59                       0.44   \n",
              "2                         0.51                       0.50   \n",
              "3                         0.43                       0.71   \n",
              "4                         0.59                       0.44   \n",
              "5                         0.53                       0.51   \n",
              "6                        -0.14                       0.42   \n",
              "7                        -0.01                      -0.13   \n",
              "8                        -0.07                       0.06   \n",
              "9                         0.50                       0.75   \n",
              "10                        0.60                       0.44   \n",
              "11                        0.57                       0.53   \n",
              "12                        0.37                       0.69   \n",
              "13                        0.59                       0.44   \n",
              "14                        0.51                       0.50   \n",
              "\n",
              "    External_dataset_gse55296  \n",
              "0                        0.43  \n",
              "1                        0.52  \n",
              "2                        0.48  \n",
              "3                        0.51  \n",
              "4                        0.54  \n",
              "5                        0.55  \n",
              "6                        0.13  \n",
              "7                        0.10  \n",
              "8                        0.18  \n",
              "9                        0.61  \n",
              "10                       0.57  \n",
              "11                       0.66  \n",
              "12                       0.43  \n",
              "13                       0.52  \n",
              "14                       0.48  "
            ]
          },
          "execution_count": 158,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Sort dataframe by performance metric\n",
        "results_bestMLmodel_ig.sort_values(by=['Performance_metric']).reset_index(drop=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TuPnx5mWeizm"
      },
      "source": [
        "---"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.11"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}