notebooks/fig3-cancer-risk.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7705a516-1b71-4740-ac38-d30b59e93fca",
   "metadata": {},
   "outputs": [],
   "source": [
    "#data\n",
    "import pandas as pd\n",
    "import os\n",
    "import statsmodels.stats.multitest as multi\n",
    "import numpy as np\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1c4132a2-e957-49e6-8395-0b93f4c696b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plotting\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import matplotlib\n",
    "matplotlib.rcParams['pdf.fonttype'] = 42\n",
    "matplotlib.rcParams['ps.fonttype'] = 42"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "de4f1028-11c9-4edd-9224-bb5b2c2e641b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#statistics\n",
    "from scipy.stats import mannwhitneyu\n",
    "import statsmodels.formula.api as smf\n",
    "import statsmodels.api as sm\n",
    "from scipy.stats import mannwhitneyu\n",
    "from scipy import stats\n",
    "from sklearn import preprocessing\n",
    "#model\n",
    "from sklearn.utils import resample\n",
    "from sklearn.linear_model import Lasso\n",
    "import sklearn.metrics as metrics\n",
    "import xgboost\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedKFold"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8bb59855-b443-4414-a95b-f68ab14591f9",
   "metadata": {},
   "source": [
    "### Get UKBB risk variants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "974cd661-1c21-4ce9-8203-47492481af10",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('/cellar/users/mpagadal/resources/tcga/ensembl_map.json', 'r') as f:\n",
    "    ensembl = json.load(f)\n",
    "ensembl={k.split(\".\")[0]:v for k,v in ensembl.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "be19db0c-8b5b-4ee0-96a3-45f8ba701a5a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_ukbb_prs(raw,sumstats,snps):\n",
    "    \n",
    "    print(\"{} PRS snps\".format(len(snps)))\n",
    "    \n",
    "    #compile raw files\n",
    "    compiled_raw=pd.read_csv(raw,delim_whitespace=True)\n",
    "    \n",
    "    #keep snps in list\n",
    "    keep_snps=[]\n",
    "    \n",
    "    for x in snps:\n",
    "        try:\n",
    "            keep_snps.append([s for s in compiled_raw.columns if x in s][0])\n",
    "        except:\n",
    "            try:\n",
    "                keep_snps.append([s for s in compiled_raw.columns if x.rsplit(\":\",2)[0]+\":\"+x.split(\":\")[3]+\":\"+x.split(\":\")[2] in s][0])\n",
    "            except:\n",
    "                pass\n",
    "    \n",
    "    compiled_raw=compiled_raw[[\"IID\"]+keep_snps]\n",
    "    \n",
    "    #format raw files\n",
    "    compiled_raw=compiled_raw.set_index(\"IID\")\n",
    "    compiled_raw=compiled_raw.T\n",
    "    compiled_raw[\"min_allele\"]=compiled_raw.index.str.split(\"_\").str[-1]\n",
    "    \n",
    "    #remove duplicates\n",
    "    compiled_raw[\"snp\"]=compiled_raw.index.str.split(\"_\").str[0]\n",
    "    compiled_raw=compiled_raw[~compiled_raw[\"snp\"].duplicated()]\n",
    "    del compiled_raw[\"snp\"]\n",
    "    \n",
    "    compiled_raw.index=compiled_raw.index.str.rsplit(\":\",2).str[0]\n",
    "    \n",
    "    #reformat summary\n",
    "    summary_merge=sumstats[[\"ID\",\"A1\",\"OR\"]].copy()\n",
    "    summary_merge=summary_merge.drop_duplicates(subset=\"ID\",keep=\"first\")\n",
    "    summary_merge=summary_merge.set_index(\"ID\")\n",
    "    summary_merge.index=summary_merge.index.str.rsplit(\":\",2).str[0]\n",
    "    \n",
    "    #annotate the genotypes\n",
    "    compiled_raw=pd.merge(compiled_raw,summary_merge, left_index=True,right_index=True,how=\"left\")\n",
    "    print(\"{} snps missing weight\".format(len(compiled_raw[compiled_raw[\"OR\"].isnull()])))\n",
    "    \n",
    "    compiled_raw[\"logOR\"]=np.log2(compiled_raw[\"OR\"])\n",
    "    del compiled_raw[\"OR\"]\n",
    "    compiled_raw=compiled_raw[~compiled_raw[\"logOR\"].isnull()] #remove null weights\n",
    "    compiled_raw[\"logOR\"]=np.where(compiled_raw[\"min_allele\"]!=compiled_raw[\"A1\"],compiled_raw[\"logOR\"]*-1,compiled_raw[\"logOR\"]) #orient the odds ratio\n",
    "    print(\"switching {} genotypes\".format(len(compiled_raw[compiled_raw[\"min_allele\"]!=compiled_raw[\"A1\"]]))) #number of mismatched genotypes\n",
    "\n",
    "    del compiled_raw[\"min_allele\"]\n",
    "    del compiled_raw[\"A1\"]\n",
    "\n",
    "    #get beta values\n",
    "    mp_beta=dict(zip(compiled_raw.index,compiled_raw[\"logOR\"]))\n",
    "    compiled_raw_switch=compiled_raw.iloc[:,:-1].T\n",
    "    \n",
    "    #align to all positive snps\n",
    "    switch_snps=[k for k,v in mp_beta.items() if v < 0]\n",
    "    \n",
    "    compiled_raw_switch=compiled_raw_switch.fillna(0)\n",
    "    \n",
    "    for x in switch_snps:\n",
    "        compiled_raw_switch[x]=2-compiled_raw_switch[x]\n",
    "        \n",
    "    compiled_raw_switch=compiled_raw_switch.T\n",
    "    compiled_raw_switch[\"beta\"]=compiled_raw_switch.index.map(mp_beta)\n",
    "    compiled_raw_switch[\"beta\"]=abs(compiled_raw_switch[\"beta\"])\n",
    "    compiled_raw_switch=compiled_raw_switch.astype(float)\n",
    "    \n",
    "    #compile weighted genotypes\n",
    "    compiled_raw_switch_out=compiled_raw_switch.iloc[:,:-1].mul(compiled_raw_switch[\"beta\"],axis=0)\n",
    "    raw_prs=compiled_raw_switch.iloc[:,:-1].mul(compiled_raw_switch[\"beta\"],axis=0).sum(axis=0).reset_index()\n",
    "    raw_prs.columns=[\"IID\",\"PRS\"]\n",
    "    \n",
    "    return(raw_prs,compiled_raw_switch_out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "7f1aaa13-bcb7-479e-96f7-9cb28ff63492",
   "metadata": {},
   "outputs": [],
   "source": [
    "def quantile_plot(df,form,num,lin_log):\n",
    "    '''\n",
    "    df: dataframe with PRS and pheno/cov in form\n",
    "    prs: prs column\n",
    "    form: formula for analysis\n",
    "    num: number of quantiles\n",
    "    '''\n",
    "    \n",
    "    if lin_log == \"logistic\":\n",
    "        \n",
    "        quant=[]\n",
    "        odds=[]\n",
    "        se=[]\n",
    "\n",
    "        for i in range(num):\n",
    "            quant.append(str((100/num)*i))\n",
    "    \n",
    "            #make groupings for logistic regression\n",
    "            \n",
    "            cases=df[(df[\"PRS\"]>df[\"PRS\"].quantile((1/num)*i))&(df[\"PRS\"]<df[\"PRS\"].quantile((1/num)*(i+1)))][\"FID\"].tolist()\n",
    "            \n",
    "            controls=df[(df[\"PRS\"]>df[\"PRS\"].quantile(int(num/2)*(1/num)))&(df[\"PRS\"]<df[\"PRS\"].quantile(int(num/2)*(1/num)+(1/(num))))][\"FID\"].tolist()\n",
    "    \n",
    "            df_filt=df[df[\"FID\"].isin(cases+controls)]\n",
    "            #df_filt=df[df[\"FID\"].isin(cases)]\n",
    "\n",
    "            df_filt[\"group\"]=np.where(df_filt[\"FID\"].isin(cases),1,0)\n",
    "            print(df_filt[\"group\"].value_counts())\n",
    "            \n",
    "            if i != int(num/2):\n",
    "                model=smf.logit(form,data=df_filt).fit()\n",
    "                odds.append(np.exp(model.params.tolist()[1]))\n",
    "                se.append(model.bse.tolist()[1])\n",
    "            else:\n",
    "                odds.append(1)\n",
    "                se.append(0)\n",
    "                \n",
    "        df_odds=pd.DataFrame({\"quant\":quant,\"odds\":odds,\"se\":se})\n",
    "        return(df_odds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "96ef7f65-b421-4050-980d-93ddd4d59f18",
   "metadata": {},
   "outputs": [],
   "source": [
    "snps=pd.read_csv(\"../tables/time-germline.csv\",delimiter=\"\\t\")\n",
    "snps[\"snp_noallele\"]=snps[\"proxy_variant\"].str.rsplit(\":\",2).str[0]\n",
    "mp_variant=dict(zip(snps[\"proxy_snp\"],snps[\"proxy_variant\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "5fbcef1e-8e75-4622-aedd-0e13d4c13de5",
   "metadata": {},
   "outputs": [],
   "source": [
    "ukbb=pd.read_csv(\"../tables/supplementary_table_10.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "dad9694e-7a38-46b8-8270-2e8b375dd4c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "fdr_thresh=0.05"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b9f0d3d8-dad1-4167-876d-4f9f0d7dc84b",
   "metadata": {},
   "outputs": [],
   "source": [
    "ukbb_sig=ukbb[ukbb[\"Overall_Pval\"]<fdr_thresh]\n",
    "ukbb_sig=ukbb_sig[ukbb_sig[\"Outcome\"].str.startswith(\"C\")]\n",
    "ukbb_annot=pd.merge(ukbb_sig, snps, on=\"proxy_variant\", how=\"left\")\n",
    "pheno_group=pd.read_csv(\"/cellar/users/mpagadal/projects/germline-immune/data/networkx/networkx.group.csv\")\n",
    "pheno_group[\"nodes\"]=pheno_group[\"nodes\"].str.replace(\".\",\"-\")\n",
    "mp_group=dict(zip(pheno_group[\"nodes\"],pheno_group[\"group\"]))\n",
    "ukbb_annot[\"group\"]=ukbb_annot[\"file\"].map(mp_group)\n",
    "ukbb_sig[\"proxy_variant\"]=ukbb_sig[\"Var1_ID\"].map(mp_variant).fillna(ukbb_sig[\"Var1_ID\"])\n",
    "\n",
    "#annotate dataframe\n",
    "ukbb_sig_snps=ukbb_sig[\"proxy_variant\"].tolist()+[x.rsplit(\":\",2)[0]+\":\"+x.split(\":\")[3]+\":\"+x.split(\":\")[2] for x in ukbb_sig[\"proxy_variant\"].tolist()]\n",
    "snps[\"ukbb\"]=np.where(snps[\"proxy_variant\"].isin(ukbb_sig_snps),\"ukbb\",\"no\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fa31ddad-da51-4a94-90b3-e3aaff175276",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "619 UKBB associations from 112 IP components\n"
     ]
    }
   ],
   "source": [
    "print(\"{} UKBB associations from {} IP components\".format(len(snps[snps[\"ukbb\"]==\"ukbb\"][\"proxy_variant\"].unique()),len(snps[snps[\"ukbb\"]==\"ukbb\"][\"file\"].unique())))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a78f69ab-8b8f-4489-a04e-8f3efb858267",
   "metadata": {},
   "outputs": [],
   "source": [
    "#get significant UKBB associations \n",
    "    #filter for cancer associations: Outcome ICD10 codes - C\n",
    "    #Adjusted for pval instead of FDR thresh\n",
    "    \n",
    "ukbb_sig=ukbb[ukbb[\"Overall_Pval\"]<fdr_thresh]\n",
    "ukbb_sig=ukbb_sig[ukbb_sig[\"Outcome\"].str.startswith(\"C\")]\n",
    "ukbb_sig[\"proxy_variant\"]=ukbb_sig[\"Var1_ID\"].map(mp_variant).fillna(ukbb_sig[\"Var1_ID\"])\n",
    "\n",
    "#annotate dataframe\n",
    "ukbb_sig_snps=ukbb_sig[\"proxy_variant\"].tolist()+[x.rsplit(\":\",2)[0]+\":\"+x.split(\":\")[3]+\":\"+x.split(\":\")[2] for x in ukbb_sig[\"proxy_variant\"].tolist()]\n",
    "snps[\"ukbb\"]=np.where(snps[\"proxy_variant\"].isin(ukbb_sig_snps),\"ukbb\",\"no\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e66b83c1-518c-48de-b367-3eb412ed463d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "660 UKBB associations from 112 IP components\n"
     ]
    }
   ],
   "source": [
    "print(\"{} UKBB associations from {} IP components\".format(len(snps[snps[\"ukbb\"]==\"ukbb\"]),len(snps[snps[\"ukbb\"]==\"ukbb\"][\"file\"].unique())))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "846bcd2c-bcc5-4a69-bcee-288d33f7190c",
   "metadata": {},
   "outputs": [],
   "source": [
    "ukbb_annot=pd.merge(ukbb_sig, snps, on=\"proxy_variant\", how=\"left\")\n",
    "ukbb_annot[\"ukbb_a1\"]=ukbb_annot[\"Var1_MAF\"].str.split(\":\").str[0]\n",
    "ukbb_annot[\"beta_match\"]=np.where(ukbb_annot['ukbb_a1']!=ukbb_annot[\"A1\"],ukbb_annot[\"beta\"]*-1,ukbb_annot[\"beta\"])\n",
    "ukbb_annot[\"logOR\"]=np.log10(ukbb_annot[\"Var1_OR\"])\n",
    "ukbb_annot[\"ref\"]=ukbb_annot[\"proxy_variant\"].str.split(\":\").str[2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "21bb4b30-3453-4095-b4a5-7ddd5ce51277",
   "metadata": {},
   "outputs": [],
   "source": [
    "pheno_group=pd.read_csv(\"/cellar/users/mpagadal/projects/germline-immune/data/networkx/networkx.group.csv\")\n",
    "pheno_group[\"nodes\"]=pheno_group[\"nodes\"].str.replace(\".\",\"-\")\n",
    "mp_group=dict(zip(pheno_group[\"nodes\"],pheno_group[\"group\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "cf6b5c86-72b4-464b-a5a4-88dfe1584839",
   "metadata": {},
   "outputs": [],
   "source": [
    "ukbb_annot[\"group\"]=ukbb_annot[\"file\"].map(mp_group)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea0333bd-b57a-47ff-8e3f-8aabba6760b2",
   "metadata": {
    "tags": []
   },
   "source": [
    "#### group1 PRS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "b2954815-44b0-4812-a128-fa2c342755b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "group1_sumstats=ukbb_annot[ukbb_annot[\"group\"]==\"group1\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "group1_sumstats=group1_sumstats[np.logical_or(group1_sumstats[\"Outcome\"]==\"C43-C44\",group1_sumstats[\"Outcome\"]==\"C60-C63\")] #\n",
    "group1_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "cb11a57b-6735-4a1d-b44a-2269b89369b4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CTSW, FCGR3B, CTSS, ERAP1, ERAP2, LYZ, HLA-DQA2, MNDA, HLA-DQA1, HLA-DRB1, HLA-C, MICB, HLA-G, HLA-DQB2, OAS1, TAP2, HLA-DRB5, HLA-DQB1, IL2RA, HLA-H, FPR1, HLA-A, HLA-B phenotypes implicated in PRS\n"
     ]
    }
   ],
   "source": [
    "print(\"{} phenotypes implicated in PRS\".format(\", \".join(ukbb_annot[(ukbb_annot[\"group\"]==\"group1\")&(ukbb_annot[\"Outcome\"]==\"C43-C44\")][\"file\"].unique())))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "03c7f167-d59c-48ec-ae08-257d90a72c40",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "158 PRS snps\n",
      "0 snps missing weight\n",
      "switching 128 genotypes\n"
     ]
    }
   ],
   "source": [
    "group1_prs,test_ukbb_res=make_ukbb_prs(\"/cellar/users/mpagadal/projects/germline-immune/data/ukbb/time-snp/dosage-multi/ukbb_bialleleic.raw\",group1_sumstats,group1_sumstats[\"ID\"].tolist())\n",
    "icd=pd.read_csv(\"/cellar/users/mpagadal/data/ukbb/plato_input/immune_cancer_lev1_pheno.txt\",delim_whitespace=True)\n",
    "group1_prs=pd.merge(group1_prs,icd,on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "296b92b3-294c-4f7e-9b4e-b276a643c479",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.5967287409834174e-24"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cat1 = group1_prs[group1_prs[\"C43-C44\"]==0]\n",
    "cat2 = group1_prs[group1_prs[\"C43-C44\"]==1]\n",
    "mannwhitneyu(cat1[\"PRS\"], cat2[\"PRS\"])[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "977de613-73b6-4f40-b97f-f276f367883e",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    144796\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.145063\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    144163\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.149523\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    144060\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.151365\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    143642\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.154279\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    143672\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.151821\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    143522\n",
      "Name: group, dtype: int64\n",
      "1    143573\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.154011\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    143643\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.156317\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    143694\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.160174\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    144967\n",
      "0     38952\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.163967\n",
      "         Iterations 8\n"
     ]
    }
   ],
   "source": [
    "group1_prs[\"melanoma\"]=group1_prs[\"C43-C44\"]\n",
    "group1_odds=quantile_plot(group1_prs,\"melanoma ~ group + AGE + SEX + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10\",10,\"logistic\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d7f1d3db-87b9-4f6d-91a6-9d010db95a3f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAbXElEQVR4nO3dfZQcdZ3v8fdn47BMABklKDABAy7OLk8m7CjyYEBxN0F5iChEllVhYVnXXUVdwyXXo8JyXODGC3rVC+EoBhCjmI2RByHyIKASCYkTkgAOsARDJsYEuAElWRIm3/tH1UBPp6enJ5mq6kx9Xuf06e5fVdd8pqdnvlO/qvr9FBGYmVl5/VnRAczMrFguBGZmJedCYGZWci4EZmYl50JgZlZyrys6wFCNGTMmxo0bV3QMM7MdyuLFi5+NiD1rLdvhCsG4ceNYtGhR0THMzHYokn430DJ3DZmZlZwLgZlZybkQmJmVnAuBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZybkQmJntAKbOXMDUmQsy2bYLgZlZybkQmJmVnAuBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZybkQmJmVnAuBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZybkQmJmVnAuBmVmTm9fVQ9fK9Ty44nmOvuwe5nX1DOv2XQjMzJrYvK4eps9dxqbeLQD0rN/I9LnLhrUYuBCYmTWxGfO72bi5t1/bxs29zJjfPWxfI7NCIOlaSWslLR9g+ZmSlqa3ByS9PassZmY7qtXrNw6pfVtkuUcwC5hcZ/kK4NiIOAy4BLgmwyxmZjukfdpah9S+LTIrBBFxP/B8neUPRMT/S5/+GhibVRYzsx3VtEkdtLaM6tfW2jKKaZM6hu1rvG7YtrR9zgFuH2ihpPOA8wD222+/vDKZmRVuyoR2AC6Ys5RNvVtob2tl2qSOV9uHQ+GFQNJ7SArBMQOtExHXkHYddXZ2Rk7RzMyawpQJ7cxeuBKAH/7TkcO+/UILgaTDgG8DJ0TEc0VmMTMrq8JOH5W0HzAX+GhEPF5UDjOzsstsj0DSbOA4YIykVcCXgRaAiLga+BKwB/B/JQG8EhGdWeUxM7PaMisEEXHGIMvPBc7N6uubmVljfGWxmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZybkQmJnVMXXmAqbOXFB0jEy5EJiZlZwLgZlZyRU+6JyZmQ0ui8Hm+niPwMys5FwIzMxKzoXAzKzkXAjMzErOhcDMrORcCMzMSs6FwMys5FwIzMxKzoXAzKzkXAjMzErOhcDMbADzunroWrmeB1c8z9GX3cO8rp6iI2XChcDMrIZ5XT1Mn7uMTb1bAOhZv5Hpc5eNyGLgQmBmVsOM+d1s3Nzbr23j5l5mzO8uKFF2XAjMzGpYvX7jkNp3ZC4EZmY17NPWOqT2HZkLgZlZDdMmddDaMqpfW2vLKKZN6igoUXY8MY2ZWQ1TJrQDcMGcpWzq3UJ7WyvTJnW82j6SuBCYmQ1gyoR2Zi9cCWQ7Q1jR3DVkZlZyLgRmZiXnQmBmVnIuBGZmJedCYGZWckMqBJLelFUQMzMrxoCnj0p6Y3UTsFDSBEAR8XymyczMLBf1riN4FvhdVVs78BsggAOyCmVmZvmp1zV0AdANnBwR+0fE/sCq9LGLgJnZCDFgIYiIrwLnAl+SdIWk3Uj2BMzMbASpe7A4IlZFxGnAz4E7gdG5pDKz0ps6cwFTZy4oOkYpNDTWUETcIuku4K0Z5zEzs5w1fPpoRGwkOW7QEEnXSlorafkAy/9S0gJJL0v6fKPbNTOz4VXv9NGbq5uA90hqA4iIkwfZ9izgm8D1Ayx/Hvg0MKWhpGZmBRjJo472qdc1NBZ4FPg2yUFiAZ3A/25kwxFxv6RxdZavBdZK+kCjYc3MbPjV6xrqBBYDXwBeiIh7gY0RcV9E3JdHuD6SzpO0SNKidevW5fmlzcxGvAH3CCJiC3ClpB+l93+ot36WIuIa4BqAzs5On8JqZjaMBv3DHhGrgNPSLpwXs49kZmZ5qnew+B3AmIi4HSAibgNuk3Qy0BMRi3PKaGZmGaq3RzADOKtG+6Mk3TTvrbdhSbOB44AxklYBXwZaACLiakl7AYuA1wNbJH0GOCgivNdhZpajeoVgj4h4uroxIp6UtMdgG46IMwZZvobkzCQzMytQvbOGWuss22W4g5iZWTHqFYK7JH1FkiobJV0M3JNtLDMzy0u9rqF/I7mY7ElJS9K28cBDJKOSmpnZCFDvOoKXgDMkHQAcnDY/EhFP5ZLMzMxyMeigcxHxVETcEhG3AB/LIZOZldy8rh66Vq7nwRXPc/Rl9zCvq6foSCPakCavBwYbaM7MbLvM6+ph+txlbOrdAkDP+o1Mn7vMxSBDQy0EGnwVM7NtN2N+Nxs39/Zr27i5lxnzuwtKNPINtRD8dSYpzMxSq9dvHFK7bb8BC4ESp0s6LX18PPA1SZ+UNNQCYlaTpyO0avu01b6EaaB22371/qB/Czgd+ChwA/AJkiEhJgJXZh/NzMpo2qQOWltG9WtrbRnFtEkdBSUa+epdR/DuiDhUUguwBtg7IjZJ+j7QlU88MyubKRPaAbhgzlI29W6hva2VaZM6Xm234VevELwCEBGbJT0UEZvS569I6q3zOjPbgfV11RU5ReOUCe3MXriy8BxlUa9raI2kXQEiYnJfYzpq6Kasg5mZWT7qXVl8wgCL/gicmE0cMzPL25DO/pH0yYh4KZ143szMRoB6M5R9rroJmC5pZ4CIuCLLYGZmlo96ewQXA0cAuwK7pfej0se7ZR/NzMzyUO+soYOBK0gmobk4IjZI+nhEXJxPNDMzy8OAewQRsTIiPgw8ANwp6cP5xTIzs7w0Mgz1T4C/IekmWpV5IjMzy1W9g8V7AntGxKMRsQGYlrYfDKyNiHU5ZTQzswzV2yP4BrBnjfaxwNeziWNmZnmrVwgOjYj7qhsjYj5wWHaRzMwsT/UKQcs2LjMzsx1IvULwhKT3VzdKOgHwBPa23TwvrVlzqHcdwWeBWyWdDixO2zqBI/FYQ7adBpqXFvBwwwZ41NE81buO4HHgUOA+YFx6uw84LF1mts08L61Z86i3R0BEvAx8N6csViKel7Y59XXXberdwtGX3eMJYUrCcw9bITwvbfMZqLvOx25GPhcCK0SzzUs7deaCV2fmKit315XXgIVA0t3p/eX5xbGymDKhnUtPPZSdRiUfwfa2Vi499VB3QxTI3XXlVe8Ywd6SjgVOlvQDkvkIXhURv8k0mY14npe2uezT1kpPjT/67q4b+eoVgi8BF5IMKVE9CU0A780qlJnlb9qkDqbPXdave6jI7jrLT705i+cAcyR9MSIuyTGT5aCvP9z/iVufvm65C+YsZVPvFtrbWn3WUEnUPX0UICIukXQyMDFtujcibs02lpkVwd115TToWUOSLgXOBx5Nb+enbWZmNgIMukcAfAAYHxFbACRdB3QB07MMZmZm+Wj0OoK2ise7ZxHEzMyK0cgewaVAl6Sfk5xCOhHvDZiZjRiNHCyeLele4B0kheB/RMSawV4n6VqSUUrXRsQhNZaLZKaz9wMbgLN8bYKV2byuHmbM72b1+o3s4zN2LEeN7BEQEb8Hbh7itmcB3wSuH2D5CcCB6e0I4Kr03qx0+sb56TuH38NyW54aKgTbIiLulzSuziqnANdHRAC/ltQmae+06AzoueeeY9asWf3aDj74YN7xjnewefNmbrzxxq1eM378eMaPH8+GDRu46aabtlre2dnJIYccwgsvvMCPf/zjrZYfeeSRdHR08Oyzz3LrrVufOTtx4kQOOOAA1qxZwx133LHV8uOPP559992XZ555hrvvvnur5ZMnT2avvfbiqaee4v77799q+YknnsiYMWPo7u5mwYKtx8P54Ac/yO67787y5ctZtGjRVstPP/10Ro8ezZIlS1iyZAkAb3nuRQBmzermzDPPpKWlhYceeohHHnlkq9efddZZADzwwAM8/nj/EchbWlo488wzAbjvvvtYsWJFv+WjR4/m9NNPB+Cuu+5i1apV/Zbvs34Tq9sOBeCOO+5gzZr+O5t77LEHJ510EgC33HILzz33XL/le+21F5MnTwZg7ty5vPjii/2Wjx07lve9730A3HTTTWzYsKHf8v33359jjz0WgP2eX8ysWf3H1Xnb297GUUcdBbDV5w6G77N3+e2P1Rzn5z9uXc6UCe25fvYqPxtZfPYqFfnZe/3rX8+pp54KFP/Zu/HGG9m8eXO/5Vl/9ioVOehcO/BMxfNVadtWJJ0naZGkRdVvltlIsObFl2u2r3vplZyTWBkp+Ye8zgrSW4FVEfGypONIJq6/PiLWD7rxZI/g1gGOEdwGXBoRv0yf3w1cEBGLq9et1NnZGbX+87ChaZYri50jcfRl99Qc56e9rZVfXZjvaC5FvxeWDUmLI6Kz1rJG9gj+E+iV9BfAd4D9ge8PQ65VwL4Vz8cCq4dhu2Y7nGYbltvKpZFjBFsi4hVJHwS+FhHfkNQ1DF/7ZuBf05FNjwBeGOz4gI08zfBfZzPMytX39XzWkBWhkUKwWdIZwMeBk9K2lsFeJGk2cBwwRtIq4Mt9r4uIq4Gfkpw6+iTJ6aNnDzW82fYaaFYuyP9snSkT2v2H3wrRSCE4G/gE8JWIWCFpf+B7g70oIs4YZHkA/9JQSrOM1JuVy3+UrSwauaDsUeDTFc9XAJdlGcosL56Va2vN0F1n+RqwEEhaRjIBTU0RcVgmiSxzzdAn3iw8K5dZ/T2CE9P7vu6bG9L7M0n69G0H1Ex94s3As3KZ1Tl9NCJ+FxG/A46OiAsiYll6uxCYlF9EG071+sTLaMqEdi499VB2GpX8KrS3tXLpqYeWsihaeTVysHgXScdUXPh1FLBLtrEsK+4T35pn5bKya6QQnANcK6lvHoL1wD9kF8my5D5xM6s26JXFEbE4It5OMrTE2yNivIeL3nH5ClYzq1bvrKHPDdAOQERckVGmEa3ocVz6+r4vmLOUTb1baPcVrGalV69raLf0voNkUpq++QhOArYeK9l2GO4TN7NKAxaCiLgYQNLPgMMj4o/p84uAH+WSzszMMtfI6KP7AZsqnm8CxmWSxszMctfIWUM3AAsl9U3dNQW4LrtIZmaWp0bGGvqKpNuBd5MMOXF2RAzHMNRmZtYEGp2zuBfYQlIItmQXx8zM8jboMQJJ5wM3AmOANwHfk/SprIOZmVk+Gr2y+IiIeAlA0uXAAuAbWQYzM7N8NHLWkEi6hvr0pm1mZjYCNLJH8F3gwaqzhr6TXSQzM8tTI2cNXSHpXuAYkj0BnzVkI46vsLYya+isoXSQOQ80Z2Y2AjVyjMDMzEYwFwIzs5Jr9IIyGwbNNGm8+8TNrI/3CHIy0KTx87p6Ck5mZmXnQpATTxpvZs3KhSAnnjTezJqVC0FOBpoc3pPGm1nRXAhy4knjzaxZ+ayhnHjSeDNrVi4EOfKk8WbWjNw1ZGZWci4EZmYl50JgZlZyLgRmZiXnQmBmVnIuBGZmJedCYGZWcqUpBFNnLmDqzAVFxzAzazqlKQRmZlZbpoVA0mRJ3ZKelHRhjeVvkPRjSUslLZR0SJZ5zMxsa5kVAkmjgG8BJwAHAWdIOqhqtf8JLImIw4CPAV/PKo+ZmdWW5R7BO4EnI+KpiNgE/AA4pWqdg4C7ASLit8A4SW/OMJOZmVXJshC0A89UPF+VtlV6GDgVQNI7gbcAY6s3JOk8SYskLVq3bl1Gcc3MyinLQqAabVH1/DLgDZKWAJ8CuoBXtnpRxDUR0RkRnXvuuefwJzUzK7Esh6FeBexb8XwssLpyhYh4ETgbQJKAFeltxPLw02bWbLLcI3gIOFDS/pJ2Aj4C3Fy5gqS2dBnAucD9aXEwM7OcZLZHEBGvSPpXYD4wCrg2Ih6R9Il0+dXAXwHXS+oFHgXOySqPmZnVlukMZRHxU+CnVW1XVzxeAByYZQYzM6vPVxabmZWcC4GZWcm5EJiZlZwLgZlZybkQmJmVXCkKwbyuHrpWrufBFc9z9GX3MK+rp+hIZmZNY8QXgnldPUyfu4xNvVsA6Fm/kelzl7kYmJmlRnwhmDG/m42be/u1bdzcy4z53QUlMjNrLiO+EKxev3FI7WZmZTPiC8E+ba1DajczK5sRXwimTeqgtWVUv7bWllFMm9RRUCIzs+aS6VhDzWDKhGQunAvmLGVT7xba21qZNqnj1XYzs7Ib8YUAkmIwe+FKwPMBmJlVG/FdQ2ZmVp8LgZlZybkQmJmVnAuBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZybkQmJmVnAuBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZyZViGGrw8NNmZgPxHoGZWcm5EJiZlZwLgZlZybkQmJmVnAuBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZySkiis4wJJLWAb/bxpePAZ4dxjjbyjn6c47+miFHM2QA56i2PTneEhF71lqwwxWC7SFpUUR0OodzOEfzZ3CO/HK4a8jMrORcCMzMSq5sheCaogOknKM/5+ivGXI0QwZwjmqZ5CjVMQIzM9ta2fYIzMysiguBmVnJjchCIGmypG5JT0q6sMZySfo/6fKlkg4fpq97raS1kpZXtL1R0p2Snkjv37AtmYeYY19JP5f0mKRHJJ2fdxZJO0taKOnhNMPFeWeo2t4oSV2Sbi0qh6SnJS2TtETSogJztEmaI+m36WfkyLxzSOpI34e+24uSPlNAjs+mn8/lkmann9sifibnpxkekfSZtC2/HBExom7AKOC/gAOAnYCHgYOq1nk/cDsg4F3Ag8P0tScChwPLK9r+F3Bh+vhC4PJtyTzEHHsDh6ePdwMeBw7KM0v63u6aPm4BHkzf69zfj3SbnwO+D9xa4M/laWBMVVsROa4Dzk0f7wS0FfVzqdjuGuAtOX9G24EVQGv6/CbgrLzfC+AQYDkwmmT64LuAA3N9L7b3h9hsN+BIYH7F8+nA9Kp1ZgJnVDzvBvYepq8/jv6F4NVtk/yB7t6WzNuZ6SfA3xSVJf2A/wY4oogMwFjgbuC9vFYIisjxNFsXglxzAK8n+eOnInNUbfdvgV/lnYOkEDwDvJHkD/CtaZa8fyanAd+ueP5F4II8c4zErqG+H26fVWnbUNcZLm+OiN8DpPdvqrFOZnkkjQMmkPxHnmuWtDtmCbAWuDMics+Q+hrJL9aWirYicgTwM0mLJZ1XUI4DgHXAd9Ousm9L2qWAHJU+AsxOH+eWIyJ6gK8CK4HfAy9ExM/yzJBaDkyUtIek0SQ9FvvmmWMkFgLVaKs+R7aRdfKUSR5JuwL/CXwmIl7MO0tE9EbEeJL/yN8p6ZC8M0g6EVgbEYu35eXDlSN1dEQcDpwA/IukiQXkeB1J9+VVETEBeImk2yHvHMkGpZ2Ak4EfDeVlw5Ej7XM/Bdgf2AfYRdLf55kBICIeAy4H7gTuIOneeSXPHCOxEKwiqaZ9xgKrt2Gd4fIHSXsDpPdra6wz7HkktZAUgRsjYm6RWSJiPXAvMLmADEcDJ0t6GvgB8F5J3ysgBxGxOr1fC/wYeGcBOVYBq9K9M4A5JIWhkM8GSVH8TUT8IX2eZ473ASsiYl1EbAbmAkflnAGAiPhORBweEROB54En8swxEgvBQ8CBkvZP/9v4CHBz1To3Ax9T4l0ku4S/zyjPzcDH08cfJ+mv35bMDZMk4DvAYxFxRRFZJO0pqS193EryS/fbPDMARMT0iBgbEePS7dwTEX+fdw5Ju0jare8xSV/08rxzRMQa4BlJHWnT8cCjeeeocAavdQuRc46VwLskjU5/Z44HHss5AwCS3pTe7wecSvKe5Jdjew/0NOONpI/tcZKj6V9I2z4BfCJ9LOBb6fJlQOcwfd3ZJH2Nm0kq9TnAHiQHKp9I79+YrrsP8NN6mbcjxzEku4dLgSXp7f15ZgEOA7rSDMuBL6Xtub8fFds8jtcOFueag6Rv/uH09kjF57KIz8d4YFH6s5kHvKGgHKOB54DdK9ry/rlcTPIPynLgBuDPC3ovfkFSkB8Gjs/7vfAQE2ZmJTcSu4bMzGwIXAjMzErOhcDMrORcCMzMSs6FwMys5FwIrNRUY8TYGutcJKlHySiZyyWdXKP9UUlnVLzmXZIeTJc9JumiHL6dWtnbJH2y4vk+kuakj49TOhqrlZsLgTU9Sa/LcPOzSK54HsyVkQyXcRpwraQ/q2o/BZiZXtENyQif56XLDiEZ2bIIbcCrhSAiVkfEhwvKYk3KhcAKJemLSsbFv1PJePCfT9vvlfQfku4Dzpd0fDpI2rL0v/g/T9d7WtKY9HGnpHvTxxdJukHSPUrGc//HWl8/Iu4nuaS/IZGMC/MKMKaq/QlgA8nFWZAMENY3YFhvRDxa43tvlfQDJXNi/DDdg+hMl/2pYr0PS5qVPj4pXa9L0l2S3lzx/V6bvm9PSfp0+vLLgLemeyYzJI2rtfeTXvl8raSH0m2f0uh7Yju+LP/TMqsr/aP3IZLRUV9HMlR15eBwbRFxrKSdSa6uPD4iHpd0PfDPJKOK1nMYyRwIuwBdkm6LdLyf7ch8BMkopuuq2g8HnohkHCGAK4HutDDdAVwXEf9dtbl/BjZExGGSDiP5/gfzS+BdERGSziUZVfXf0mV/CbyHZA6KbklXkQwod0i6Z9I3Gm0tXyAZfuMf0qFBFkq6KyJeaiCT7eC8R2BFOgb4SURsjIg/ArdULf9het9BMjjY4+nz60gmARpM37afBX5OMsjbtvqskiG1vwpMjdcuyf+spG6SYb4v6ls5Iv4d6AR+BvwdSTGoNhH4Xrr+UpIhHwYzFpgvaRkwDTi4YtltEfFy+v2uBd7c+LfH3wIXpt/jvcDOwH5DeL3twFwIrEi1htCt1PffaL31XuG1z/HOVcuqx0/ZnvFUroyI8RHx7oj4RVV7BzAVuD7de0m+WMR/RcRVJIOZvV3SHjW2O1CmyvbK7+sbwDcj4lDgn6qWvVzxuJeh7fEL+FD6PY6PiP3SbjArARcCK9IvgZOUzBO7K/CBAdb7LTBO0l+kzz8K3Jc+fhr46/Txh6ped0q67T1IBpx7aLiCV4tkqO9FpKNFSvpAOqIlJNMO9gLrq152P3Bmuv4hJF1Zff4g6a/Sg9IfrGjfHehJH3+cwf2RpKtoMPOBT/VlljShgdfYCOFCYIWJiIdIhsx9mGQs+EXACzXW+2/gbOBHaZfIFuDqdPHFwNcl/YLkj22lhcBtwK+BS2odH5A0G1gAdEhaJemc7fiW/h34XPrH+6Mk/fRLSEa1PDMiqvNdBewqaSlJX//CimUXkkydeA/pQefURSTvwy+AZwcLFBHPAb9KT3udUWfVS0jmll6aHky+ZLBt28jh0UetUJJ2jYg/KZmi736SUy4bOWg62HYvAv4UEV/d3m3lJT2w/PmIWFR0FisXnzVkRbtG0kEkfd3XDUcRMLOh8R6BmVnJ+RiBmVnJuRCYmZWcC4GZWcm5EJiZlZwLgZlZyf1/uL5Z3BfUumoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.errorbar(x=group1_odds[\"quant\"].tolist(),y=group1_odds[\"odds\"].tolist(),yerr=group1_odds[\"se\"].tolist(),fmt=\"o\")\n",
    "plt.xlabel(\"group 1 PRS quantile\")\n",
    "plt.ylabel(\"odds of C43-C44\")\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.savefig(\"../plots/fig3/group1.melanoma.ukbb.prs.quantile.pdf\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ce9255ac-4d9b-44e9-b7bf-20fced64e01b",
   "metadata": {
    "tags": []
   },
   "source": [
    "#### MHC II signature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "99117f18-eb8e-4112-bedf-557498cbd79b",
   "metadata": {},
   "outputs": [],
   "source": [
    "mhcII_sumstats=ukbb_annot[ukbb_annot[\"file\"]==\"Sigs160-Wolf_MHC2-21978456\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "mhcII_sumstats=mhcII_sumstats[mhcII_sumstats[\"Outcome\"]==\"C43-C44\"]\n",
    "mhcII_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "115699c5-8b59-465e-aaaf-dc3247aaf69a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13 PRS snps\n",
      "0 snps missing weight\n",
      "switching 13 genotypes\n"
     ]
    }
   ],
   "source": [
    "mhcII_prs=make_ukbb_prs(\"../data/ukbb/time-snp/dosage-multi/ukbb_bialleleic.raw\",mhcII_sumstats,mhcII_sumstats[\"ID\"].tolist())\n",
    "icd=pd.read_csv(\"/cellar/users/mpagadal/data/ukbb/plato_input/immune_cancer_lev1_pheno.txt\",delim_whitespace=True)\n",
    "mhcII_prs=pd.merge(mhcII_prs[0],icd,on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "16304b7b-9e63-459b-b5b8-abeeac5b633b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8.930627728989479e-15"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cat1 = mhcII_prs[mhcII_prs[\"C43-C44\"]==0]\n",
    "cat2 = mhcII_prs[mhcII_prs[\"C43-C44\"]==1]\n",
    "mannwhitneyu(cat1[\"PRS\"], cat2[\"PRS\"])[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "a5f93527-fed6-490f-ae10-da851df09213",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    145295\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.147316\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    144262\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.155072\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    142965\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.153260\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    141986\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.149079\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    134254\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.150230\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    141075\n",
      "Name: group, dtype: int64\n",
      "1    142468\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.155876\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    142989\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.159624\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    142964\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.157950\n",
      "         Iterations 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    144529\n",
      "0     36505\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.160906\n",
      "         Iterations 8\n"
     ]
    }
   ],
   "source": [
    "mhcII_prs[\"melanoma\"]=mhcII_prs[\"C43-C44\"]\n",
    "mhcII_odds=quantile_plot(mhcII_prs,\"melanoma ~ group + AGE + SEX + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10\",10,\"logistic\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "076e71e6-d796-4020-a8a9-1d948fc639b2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.errorbar(x=mhcII_odds[\"quant\"].tolist(),y=mhcII_odds[\"odds\"].tolist(),yerr=mhcII_odds[\"se\"].tolist(),fmt=\"o\")\n",
    "plt.xlabel(\"MHC II signature PRS quantile\")\n",
    "plt.ylabel(\"odds of C43-C44\")\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.savefig(\"../plots/fig3/mhcII.sig.melanoma.ukbb.prs.quantile.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "4d9b76f4-a885-4583-8c4d-b6f76c1d8e3f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f=plt.figure(figsize=(7,5))\n",
    "plt.errorbar(x=group1_odds[\"quant\"].tolist(),y=group1_odds[\"odds\"].tolist(),yerr=group1_odds[\"se\"].tolist(),fmt=\"o\",color=\"blue\",label=\"group1 genes\")\n",
    "plt.errorbar(x=mhcII_odds[\"quant\"].tolist(),y=mhcII_odds[\"odds\"].tolist(),yerr=mhcII_odds[\"se\"].tolist(),fmt=\"o\",color=\"red\",label=\"MHC II signature\")\n",
    "plt.ylabel(\"quantile\")\n",
    "plt.ylabel(\"odds of C43-C44 \\n (melanoma and other malignant neoplasms of skin)\")\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.legend()\n",
    "plt.savefig(\"../plots/fig3/mhcI.II.ukbb.quantile.pdf\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "030148f0-2dd3-4858-93ee-88f1eb39ded3",
   "metadata": {
    "tags": []
   },
   "source": [
    "### UKBB with combined MHC+group1 SNPs, Melanoma training and validation ONLY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "c3bff65e-a6cf-4512-9f99-b8403c0a2d27",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "181 PRS snps\n",
      "0 snps missing weight\n",
      "switching 148 genotypes\n"
     ]
    }
   ],
   "source": [
    "group1_sumstats=ukbb_annot[ukbb_annot[\"group\"]==\"group1\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "group1_sumstats=group1_sumstats[np.logical_or(group1_sumstats[\"Outcome\"]==\"C43-C44\",group1_sumstats[\"Outcome\"]==\"C60-C63\")] #\n",
    "group1_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "mhcII_sumstats=ukbb_annot[ukbb_annot[\"file\"]==\"Sigs160-Wolf_MHC2-21978456\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "mhcII_sumstats=mhcII_sumstats[np.logical_or(mhcII_sumstats[\"Outcome\"]==\"C43-C44\",mhcII_sumstats[\"Outcome\"]==\"C60-C63\")]\n",
    "mhcII_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "group1_mhcII_sumstats=group1_sumstats.append(mhcII_sumstats)\n",
    "group1_mhcII_sumstats\n",
    "\n",
    "group1_mhc_prs,test_ukbb_res=make_ukbb_prs(\"../data/ukbb/time-snp/dosage-multi/ukbb_bialleleic.raw\",group1_mhcII_sumstats,group1_mhcII_sumstats[\"ID\"].tolist())\n",
    "icd=pd.read_csv(\"/cellar/users/mpagadal/data/ukbb/plato_input/immune_cancer_lev1_pheno.txt\",delim_whitespace=True)\n",
    "group1_mhc_prs=pd.merge(group1_mhc_prs,icd,on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b298b4e6-5cb6-4d7c-96aa-3150aaa84f16",
   "metadata": {},
   "source": [
    "#### melanoma genes ONLY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ba7cbe4e-f822-45f5-91bd-36844b1dc37a",
   "metadata": {},
   "outputs": [],
   "source": [
    "random_sate=75 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "145f10e4-2c9b-43a5-bfce-99ceb9dbe681",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "124 PRS snps\n",
      "0 snps missing weight\n",
      "switching 115 genotypes\n"
     ]
    }
   ],
   "source": [
    "group1_sumstats=ukbb_annot[ukbb_annot[\"group\"]==\"group1\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "group1_sumstats=group1_sumstats[group1_sumstats[\"Outcome\"]==\"C43-C44\"]\n",
    "group1_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "mhcII_sumstats=ukbb_annot[ukbb_annot[\"file\"]==\"Sigs160-Wolf_MHC2-21978456\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "mhcII_sumstats=mhcII_sumstats[mhcII_sumstats[\"Outcome\"]==\"C43-C44\"]\n",
    "mhcII_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "group1_mhcII_sumstats=group1_sumstats.append(mhcII_sumstats)\n",
    "group1_mhcII_sumstats\n",
    "\n",
    "group1_mhc_prs,test_ukbb_res=make_ukbb_prs(\"../data/ukbb/time-snp/dosage-multi/ukbb_bialleleic.raw\",group1_mhcII_sumstats,group1_mhcII_sumstats[\"ID\"].tolist())\n",
    "icd=pd.read_csv(\"/cellar/users/mpagadal/data/ukbb/plato_input/immune_cancer_lev1_pheno.txt\",delim_whitespace=True)\n",
    "group1_mhc_prs=pd.merge(group1_mhc_prs,icd,on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "d4afa76a-b69b-4a68-bc80-b2b83383fe30",
   "metadata": {},
   "outputs": [],
   "source": [
    "#merge group1_mhc_prs with raw SNP data then format for model building\n",
    "group1_mhc_prs_short=group1_mhc_prs[[\"IID\",\"C43-C44\",\"C60-C63\",\"AGE\",\"SEX\",\"PC1\",\"PC2\",\"PC3\",\"PC4\",\"PC5\",\"PC6\",\"PC7\",\"PC8\",\"PC9\",\"PC10\"]]\n",
    "test_ukbb_res_t=test_ukbb_res.transpose()\n",
    "SNP_outcome_cov_table=pd.merge(test_ukbb_res_t,group1_mhc_prs,right_on=\"IID\",how=\"left\",left_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "c0a4b56f-255c-464f-8c91-dc6fbabf769a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(21701.0, 14960.0, 7291.0)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SNP_outcome_cov_table[\"Cancer\"]=SNP_outcome_cov_table[[\"C43-C44\",\"C60-C63\"]].max(axis=1)\n",
    "np.nansum(SNP_outcome_cov_table[\"Cancer\"]),np.nansum(SNP_outcome_cov_table[\"C43-C44\"]),np.nansum(SNP_outcome_cov_table[\"C60-C63\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "579899fc-e8c9-4fa5-959b-596197657d4c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Try joint gene score but slim down to melanoma only\n",
    "SNP_outcome_cov_table.columns.values\n",
    "SNP_outcome_cov_table[\"AnyCancer\"]=SNP_outcome_cov_table[['C73-C75', 'C81-C96', 'C40-C41',\n",
    "        'C50-C50',  'C00-C14', 'C45-C49',\n",
    "       'C30-C39', 'C97-C97', 'C76-C80', 'C60-C63',\n",
    "       'C15-C26', 'C51-C58', 'C43-C44', 'C69-C72', 'C64-C68']].max(axis=1)\n",
    "\n",
    "SNP_outcome_cov_table=SNP_outcome_cov_table[np.logical_or(SNP_outcome_cov_table[\"C43-C44\"]==1,SNP_outcome_cov_table[\"AnyCancer\"]==0)]\n",
    "SNP_outcome_cov_table[\"Cancer\"]=SNP_outcome_cov_table[[\"C43-C44\"]].max(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "1ae8949e-5740-44a3-9807-16ca2af084cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table=SNP_outcome_cov_table.reset_index()\n",
    "SNP_outcome_cov_table=SNP_outcome_cov_table[~SNP_outcome_cov_table[\"Cancer\"].isnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "20a87f89-d63a-4e21-81bf-eda3077da947",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table = SNP_outcome_cov_table.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "5d796d7b-4d14-45b1-a642-0e15e600b38d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(341376, 174)\n"
     ]
    }
   ],
   "source": [
    "# Separate majority and minority classes\n",
    "df_majority = SNP_outcome_cov_table[SNP_outcome_cov_table[\"Cancer\"]==0]\n",
    "df_minority = SNP_outcome_cov_table[SNP_outcome_cov_table[\"Cancer\"]==1]\n",
    " \n",
    "# Downsample majority class\n",
    "df_majority_downsampled = resample(df_majority, replace=False, random_state=random_sate, n_samples=len(df_majority))\n",
    "print(df_majority_downsampled.shape)\n",
    "\n",
    "# Combine minority class with downsampled majority class\n",
    "SNP_outcome_cov_table_downsamp = pd.concat([df_majority_downsampled, df_minority])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "08aeb26e-a696-428d-97f8-b35463b8156b",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_downsamp.drop([\"IID\",\"Cancer\",\"C43-C44\",\"C60-C63\",\"AGE\",\"SEX\",\"PC1\",\"PC2\",\"PC3\",\"PC4\",\"PC5\",\"PC6\",\"PC7\",\"PC8\",\"PC9\",\"PC10\",'PRS',\n",
    "       'FID', 'D55-D59', 'D80-D89', 'E70-E90', 'I70-I79', 'K20-K31',\n",
    "       'K50-K52', 'K70-K77', 'K80-K87', 'K90-K93', 'L40-L45', 'M00-M03',\n",
    "       'M05-M14', 'M15-M19', 'M20-M25', 'M70-M79', 'M45-M49', 'M60-M63',\n",
    "       'R70-R79', 'C73-C75', 'C81-C96', 'D60-D64', 'C40-C41', 'D37-D48',\n",
    "       'D70-D77', 'C50-C50', 'D65-D69', 'C00-C14', 'C45-C49', 'D00-D09',\n",
    "       'C30-C39', 'C97-C97', 'D50-D53', 'C76-C80', 'C15-C26', 'C51-C58',\n",
    "       'D10-D36', 'C69-C72', 'C64-C68','AnyCancer'],axis=1)\n",
    "SNP_outcome_cov_table_training_x.index=SNP_outcome_cov_table_downsamp[\"IID\"]\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_training_x.drop([\"level_0\",\"index\"],axis=1)\n",
    "SNP_outcome_cov_table_training_y=SNP_outcome_cov_table_downsamp[\"Cancer\"]\n",
    "SNP_outcome_cov_table_training_y.index=SNP_outcome_cov_table_downsamp[\"IID\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b8d89f9-effc-452f-b30d-1ff7f3bff4db",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Generating and preparing the validation datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "da126bba-4665-4a7d-a9b5-e8d8dbd53452",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "124 PRS snps\n",
      "0 snps missing weight\n",
      "switching 111 genotypes\n"
     ]
    }
   ],
   "source": [
    "geneva_prs,geneva_snps=make_ukbb_prs(\"../data/genotypes/geneva_melanoma/geneva.raw\",group1_mhcII_sumstats,group1_mhcII_sumstats[\"ID\"].tolist())\n",
    "\n",
    "pheno=pd.read_csv(\"/cellar/controlled/dbgap-genetic/phs000187v1_geneva_melanoma/phenotypes/phs000187.v1.pht000814.v1.p1.c1.Melanoma_Data.GRU.txt.gz\",comment=\"#\",delimiter=\"\\t\")\n",
    "pheno[\"case_control\"]=pheno[\"case_control\"].map({1:1,2:0,3:np.nan})\n",
    "pheno[\"pheno\"]=pheno[\"case_control\"]\n",
    "pheno=pheno.rename(columns={\"Geneva_ID_\":\"IID\"})\n",
    "pheno[\"age\"]=pheno[\"Age_at_DX\"].fillna(pheno[\"Control_Age_at_DX\"])\n",
    "geneva_prs=pd.merge(geneva_prs,pheno[[\"IID\",\"pheno\",\"age\",\"Gender_\",\"Breslow_tumor_thickness\",\"family_history\"]],on=\"IID\",how=\"left\")\n",
    "geneva_prs[\"sex\"]=geneva_prs[\"Gender_\"].map({\"M\":0,\"F\":1})\n",
    "pcs=pd.read_csv(\"/cellar/users/mpagadal/ancestry-analysis/geneva/all.cluster.mds\",delim_whitespace=True)\n",
    "pcs[\"IID\"]=pcs[\"IID\"].astype(str)\n",
    "geneva_prs[\"IID\"]=geneva_prs[\"IID\"].astype(str)\n",
    "\n",
    "geneva_prs=pd.merge(geneva_prs,pcs,on=\"IID\",how=\"left\")\n",
    "geneva_prs[\"FID\"]=geneva_prs[\"IID\"]\n",
    "\n",
    "#Merge geneva_prs file with SNP inputs, then chop up into train/test\n",
    "geneva_snps_t=geneva_snps.transpose()\n",
    "geneva_snps_t[\"IID\"]=geneva_snps_t.index\n",
    "\n",
    "geneva_snps_t = geneva_snps_t.reset_index()\n",
    "geneva_prs = geneva_prs.reset_index()\n",
    "\n",
    "geneva_snps_t[\"IID\"]=geneva_snps_t['IID'].astype(str)\n",
    "geneva_prs[\"IID\"]=geneva_prs['IID'].astype(str)\n",
    "\n",
    "geneva_outcome_cov_table=pd.merge(pd.DataFrame(geneva_snps_t),pd.DataFrame(geneva_prs),on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "f30bab8e-3e90-4a51-bead-7f1a856c6517",
   "metadata": {},
   "outputs": [],
   "source": [
    "geneva_outcome_cov_table=geneva_outcome_cov_table[~geneva_outcome_cov_table[\"pheno\"].isnull()]\n",
    "geneva_outcome_cov_table_x=geneva_outcome_cov_table.drop(['IID',\n",
    "       'PRS', 'pheno', 'age', 'Gender_', 'Breslow_tumor_thickness',\n",
    "       'family_history', 'sex', 'FID', 'SOL', 'C1', 'C2', 'C3', 'C4',\n",
    "       'C5', 'C6', 'C7', 'C8', 'C9', 'C10'],axis=1)\n",
    "geneva_outcome_cov_table_x.index=geneva_outcome_cov_table[\"IID\"]\n",
    "geneva_outcome_cov_table_x=geneva_outcome_cov_table_x.drop([\"index_x\",\"index_y\"],axis=1)\n",
    "\n",
    "geneva_outcome_cov_table_y=geneva_outcome_cov_table[\"pheno\"]\n",
    "geneva_outcome_cov_table_y.index=geneva_outcome_cov_table[\"IID\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "22a8e5f9-1ce7-42f7-a65b-13b074036b75",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table_training_y=pd.to_numeric(SNP_outcome_cov_table_training_y, errors='coerce').astype(\"Int32\")\n",
    "geneva_outcome_cov_table_y=pd.to_numeric(geneva_outcome_cov_table_y, errors='coerce').astype(\"Int32\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "015fa9db-3f83-4539-99fe-1726ed3eb262",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((356336, 115), (356336,), (3029, 115), (3029,))"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SNP_outcome_cov_table_training_x.shape,SNP_outcome_cov_table_training_y.shape,geneva_outcome_cov_table_x.shape,geneva_outcome_cov_table_y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "861fd687-0781-41df-b652-6c0123871834",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5390563113191547\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['11:65477306' '1:150456350' '11:65585257' '5:96374150' '5:96200770'\n",
      " '5:96257285' '11:65407886' '6:32438298' '6:32606941' '5:96101944'\n",
      " '6:32592605' '5:96323549' '5:96124330' '6:31239896' '6:31461771'\n",
      " '5:96057882' '1:150406172' '5:96092990' '1:150861859' '5:96115721'\n",
      " '5:96206563' '6:32581973' '6:32731258' '5:96147733' '6:32614080'\n",
      " '6:32789480' '6:32580657' '5:96128022' '5:96125330' '11:65640906'\n",
      " '11:65747057' '1:150467096' '1:150958977' '1:150635066' '1:150558293'\n",
      " '5:96185467' '5:95897963' '5:96198479' '6:29910358' '6:32636866'\n",
      " '11:65473798' '5:95999526' '5:96174446' '5:96229272' '1:150527294'\n",
      " '5:96174186' '5:96182717' '11:65583066' '1:150878649' '5:95981261'\n",
      " '6:31322980' '12:113411806' '11:65412467' '5:96120933' '5:96216387'\n",
      " '1:150737220' '5:95909680' '5:96293021' '6:32561327' '6:32610059'\n",
      " '1:150618632' '5:96151760' '6:32602396' '6:32575658' '6:32683121'\n",
      " '6:32389255' '6:32612397' '6:32592796' '6:32582194' '6:32584625'\n",
      " '6:32581582' '6:32583274']\n"
     ]
    }
   ],
   "source": [
    "#Now recreate model from fig. 7 using this new composite dataframe\n",
    "\n",
    "#filter using LASSO regression\n",
    "model = Lasso(alpha=0.000015,random_state=random_sate)  \n",
    "\n",
    "# fit model\n",
    "model.fit(SNP_outcome_cov_table_training_x, SNP_outcome_cov_table_training_y)\n",
    "\n",
    "p = model.predict(SNP_outcome_cov_table_training_x)  # test the output by changing values\n",
    "#Find model that gives best AUC, do a cutoff of weakest contributing SNPS, then run all analysis portions\n",
    "# calculate the fpr and tpr for all thresholds of the classification\n",
    "#preds = np.around(p, decimals=0, out=None)\n",
    "\n",
    "preds=p\n",
    "fpr, tpr, threshold = metrics.roc_curve(list(SNP_outcome_cov_table_training_y), preds,pos_label=1)\n",
    "roc_auc = metrics.auc(fpr, tpr)\n",
    "\n",
    "print(roc_auc)\n",
    "# method I: plt\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.title('Simple LASSO AUC')\n",
    "plt.plot(fpr, tpr, 'b', label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],'r--')\n",
    "plt.xlim([0, 1])\n",
    "plt.ylim([0, 1])\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()\n",
    "\n",
    "#now drop SNPs with low contribution\n",
    "print(SNP_outcome_cov_table_training_x.columns.values[model.coef_==0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "4abcb5ae-9e08-428a-8054-34a267d5c995",
   "metadata": {},
   "outputs": [],
   "source": [
    "#remake SNP list with dropped \"useless\" SNPs\n",
    "geneva_outcome_cov_table_dropped=geneva_outcome_cov_table.drop(SNP_outcome_cov_table_training_x.columns.values[model.coef_==0],axis=1)\n",
    "SNP_outcome_cov_table_downsamp_dropped=SNP_outcome_cov_table_downsamp.drop(SNP_outcome_cov_table_training_x.columns.values[model.coef_==0],axis=1)\n",
    "\n",
    "geneva_outcome_cov_table_x=geneva_outcome_cov_table_dropped.drop(['IID',\n",
    "       'PRS', 'pheno', 'age', 'Gender_', 'Breslow_tumor_thickness',\n",
    "       'family_history', 'sex', 'FID', 'SOL', 'C1', 'C2', 'C3', 'C4',\n",
    "       'C5', 'C6', 'C7', 'C8', 'C9', 'C10'],axis=1)\n",
    "geneva_outcome_cov_table_x.index=geneva_outcome_cov_table_dropped[\"IID\"]\n",
    "geneva_outcome_cov_table_x=geneva_outcome_cov_table_x.drop([\"index_x\",\"index_y\"],axis=1)\n",
    "\n",
    "geneva_outcome_cov_table_y=geneva_outcome_cov_table_dropped[\"pheno\"]\n",
    "geneva_outcome_cov_table_y.index=geneva_outcome_cov_table_dropped[\"IID\"]\n",
    "\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_downsamp_dropped.drop([\"IID\",\"Cancer\",\"C43-C44\",\"C60-C63\",\"AGE\",\"SEX\",\"PC1\",\"PC2\",\"PC3\",\"PC4\",\"PC5\",\"PC6\",\"PC7\",\"PC8\",\"PC9\",\"PC10\",'PRS',\n",
    "       'FID', 'D55-D59', 'D80-D89', 'E70-E90', 'I70-I79', 'K20-K31',\n",
    "       'K50-K52', 'K70-K77', 'K80-K87', 'K90-K93', 'L40-L45', 'M00-M03',\n",
    "       'M05-M14', 'M15-M19', 'M20-M25', 'M70-M79', 'M45-M49', 'M60-M63',\n",
    "       'R70-R79', 'C73-C75', 'C81-C96', 'D60-D64', 'C40-C41', 'D37-D48',\n",
    "       'D70-D77', 'C50-C50', 'D65-D69', 'C00-C14', 'C45-C49', 'D00-D09',\n",
    "       'C30-C39', 'C97-C97', 'D50-D53', 'C76-C80', 'C15-C26', 'C51-C58',\n",
    "       'D10-D36', 'C69-C72', 'C64-C68',\"AnyCancer\"],axis=1)\n",
    "SNP_outcome_cov_table_training_x.index=SNP_outcome_cov_table_downsamp_dropped[\"IID\"]\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_training_x.drop([\"level_0\",\"index\"],axis=1)\n",
    "SNP_outcome_cov_table_training_y=SNP_outcome_cov_table_downsamp_dropped[\"Cancer\"]\n",
    "SNP_outcome_cov_table_training_y.index=SNP_outcome_cov_table_downsamp_dropped[\"IID\"]\n",
    "\n",
    "geneva_outcome_cov_table_dropped=geneva_outcome_cov_table_dropped.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "b7658f37-c4d9-446c-97e7-6eaee2ff5563",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.DataFrame(SNP_outcome_cov_table_training_x.columns.values).to_csv(\"../data/risk_model/melanoma_final_snps.txt\",sep=\"\\t\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "ff75fd25-0a96-4975-9453-5bab10104efe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[09:22:26] WARNING: /tmp/abs_40obctay9q/croots/recipe/xgboost-split_1659548945886/work/src/learner.cc:576: \n",
      "Parameters: { \"verbose_eval\" } might not be used.\n",
      "\n",
      "  This could be a false alarm, with some parameters getting used by language bindings but\n",
      "  then being mistakenly passed down to XGBoost core, or some parameter actually being used\n",
      "  but getting flagged wrongly here. Please open an issue if you find any such cases.\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "regressor_xgb = xgboost.XGBRFRegressor(n_estimators = 200, random_state = random_sate,learning_rate=0.3,max_depth=9,verbose_eval=10)\n",
    "regressor_xgb.fit(SNP_outcome_cov_table_training_x, SNP_outcome_cov_table_training_y)  \n",
    "\n",
    "p = regressor_xgb.predict(geneva_outcome_cov_table_x)  # test the output by changing values\n",
    "p_disc=regressor_xgb.predict(SNP_outcome_cov_table_training_x)\n",
    "\n",
    "geneva_outcome_cov_table_dropped[\"LASSO_burden\"]=p\n",
    "SNP_outcome_cov_table_downsamp_dropped[\"LASSO_burden\"]=p_disc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "e24cd1c8-87d3-459d-b22d-47172627d839",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans.\n",
      "findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "features = SNP_outcome_cov_table_training_x.columns.values.tolist()\n",
    "features=[x.rsplit(\"_\")[0] for x in features]\n",
    "importances = regressor_xgb.feature_importances_\n",
    "\n",
    "mapping_table=pd.concat([pd.DataFrame(features),pd.DataFrame(importances)],axis=1)\n",
    "mapping_table.columns=[\"ID\",\"importance\"]\n",
    "\n",
    "rsid_table=pd.read_csv(\"../tables/time-germline.csv\",sep=\"\\t\")\n",
    "rsid_table=rsid_table.drop_duplicates(subset=\"rsid\")\n",
    "rsid_table=rsid_table[[\"snp_noallele\",\"rsid\"]]\n",
    "\n",
    "compiled_icb_rsid_plot=pd.merge(mapping_table,rsid_table,left_on=\"ID\",right_on=\"snp_noallele\")\n",
    "\n",
    "compiled_icb_rsid_plot=compiled_icb_rsid_plot.sort_values(by=\"importance\")\n",
    "compiled_icb_rsid_plot=compiled_icb_rsid_plot.iloc[compiled_icb_rsid_plot.shape[0]-15:compiled_icb_rsid_plot.shape[0],:]\n",
    "\n",
    "plt.rcParams['font.family'] = 'sans-serif'\n",
    "plt.rcParams['font.sans-serif'] = 'Helvetica'\n",
    "\n",
    "plt.rcParams['axes.edgecolor']='darkslategrey'\n",
    "plt.rcParams['axes.linewidth']=1\n",
    "plt.rcParams['xtick.color']='black'\n",
    "plt.rcParams['ytick.color']='black'\n",
    "plt.style.use('seaborn-dark-palette')\n",
    "\n",
    "my_range=list(range(0,len(compiled_icb_rsid_plot[\"importance\"])))\n",
    "fig, ax = plt.subplots(figsize=(8,10))\n",
    "plt.hlines(y=my_range, xmin=0, xmax=compiled_icb_rsid_plot[\"importance\"], alpha=0.4, linewidth=7)\n",
    "plt.plot(compiled_icb_rsid_plot[\"importance\"], my_range, \"o\", markersize=7, alpha=0.7)\n",
    "# set labels style\n",
    "ax.set_xlabel('Relative Importance', fontsize=15, fontweight='black', color = '#333F4B')\n",
    "ax.set_ylabel('')\n",
    "# change the style of the axis spines\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "# ax.spines['left'].set_bounds((0, len(my_range)-1))\n",
    "ax.set_xlim(0,0.035)\n",
    "plt.yticks(range(len(compiled_icb_rsid_plot[\"importance\"])), compiled_icb_rsid_plot[\"rsid\"])\n",
    "# add some space between the axis and the plot\n",
    "ax.spines['left'].set_position(('outward', 8))\n",
    "ax.spines['bottom'].set_position(('outward', -10))\n",
    "plt.savefig(\"../plots/fig3/melanoma_feature_importance_V2.pdf\",bbox_inches = \"tight\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "964cc0a4-44dc-4bb0-8bfe-3f000457ebb9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:6: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \n",
      "findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "geneva_outcome_cov_table2=geneva_outcome_cov_table_dropped[geneva_outcome_cov_table_dropped[\"family_history\"]!=1]\n",
    "geneva_outcome_cov_table_plot=geneva_outcome_cov_table2\n",
    "geneva_outcome_cov_table_plot=geneva_outcome_cov_table_plot.loc[geneva_outcome_cov_table_plot['LASSO_burden']<float(geneva_outcome_cov_table_plot['LASSO_burden'].quantile([0.99]))]\n",
    "minmax_scale = preprocessing.MinMaxScaler(feature_range=(0, 10))\n",
    "x = geneva_outcome_cov_table_plot[\"LASSO_burden\"].values.reshape(-1, 1)\n",
    "geneva_outcome_cov_table_plot[\"LASSO_scaled\"] = minmax_scale.fit_transform(x)\n",
    "\n",
    "fig = plt.figure(figsize=(5,5))\n",
    "\n",
    "sns.violinplot(x=\"pheno\",y=\"LASSO_scaled\",data=geneva_outcome_cov_table_plot)\n",
    "\n",
    "cat1 = geneva_outcome_cov_table2[geneva_outcome_cov_table2[\"pheno\"]==0]\n",
    "cat2 = geneva_outcome_cov_table2[geneva_outcome_cov_table2[\"pheno\"]==1]\n",
    "disc_score=cat2[\"LASSO_burden\"].mean()-cat1[\"LASSO_burden\"].mean()\n",
    "\n",
    "plt.title(stats.mannwhitneyu(cat1[\"LASSO_burden\"], cat2[\"LASSO_burden\"])[1])\n",
    "\n",
    "plt.ylabel(\"burden (pvalue < 0.05) RF Regression\")\n",
    "plt.xticks(rotation=90)\n",
    "#plt.savefig(\"/cellar/users/tsears/projects/germline-immune/plots/fig5/melanomaViolin_FH_only_RescaledOutlierDroppedFromPlot.pdf\")\n",
    "plt.savefig(\"../plots/fig3/melanomaViolin_FH_only_rescaled_99th_dropped.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "7a5665dc-b0fc-4bd6-bc97-9a899b20024d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  after removing the cwd from sys.path.\n"
     ]
    }
   ],
   "source": [
    "#restructure PRS score so that it makes more sense to plot\n",
    "geneva_outcome_cov_table_dropped.columns.values\n",
    "geneva_outcome_cov_table_dropped[\"PRS\"]=geneva_outcome_cov_table_dropped[\"LASSO_burden\"]\n",
    "geneva_outcome_cov_table2[\"PRS\"]=geneva_outcome_cov_table2[\"LASSO_burden\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "aca011fa-19e1-4b09-8dd1-4300a8f335dc",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    169\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.551226\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    170\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.519002\n",
      "         Iterations 6\n",
      "1    169\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.545223\n",
      "         Iterations 6\n",
      "1    170\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.540988\n",
      "         Iterations 6\n",
      "1    169\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.501235\n",
      "         Iterations 6\n",
      "1    169\n",
      "Name: group, dtype: int64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    170\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.529732\n",
      "         Iterations 6\n",
      "1    170\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.508096\n",
      "         Iterations 6\n",
      "1    170\n",
      "0    169\n",
      "Name: group, dtype: int64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.483214\n",
      "         Iterations 6\n",
      "1    169\n",
      "0    169\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.464374\n",
      "         Iterations 7\n"
     ]
    }
   ],
   "source": [
    "geneva_outcome_odds=quantile_plot(geneva_outcome_cov_table2,\"pheno ~ group + age + sex + C1 + C2 + C3 + C4 + C5 + C6 + C7 + C8 + C9 + C10\",10,\"logistic\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "10b7b77c-0f7f-4160-b779-a5e44cc3865f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f=plt.figure(figsize=(7,5))\n",
    "\n",
    "plt.errorbar(x=geneva_outcome_odds[\"quant\"].tolist(),y=geneva_outcome_odds[\"odds\"].tolist(),yerr=geneva_outcome_odds[\"se\"].tolist(),fmt=\"o\",color=\"blue\",label=\"group1 + mhc genes\")\n",
    "plt.ylabel(\"quantile\")\n",
    "plt.ylabel(\"odds of C43-C44 \\n (melanoma and other malignant neoplasms of skin)\")\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.savefig(\"../plots/fig3/melanoma.quantile.FH.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "e6e85aad-1154-4730-9c31-7b7540e7e9b7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#1st vs 10th quantile\n",
    "geneva_outcome_odds_filt=geneva_outcome_odds[geneva_outcome_odds[\"quant\"].isin([\"0.0\",\"90.0\"])]\n",
    "f=plt.figure(figsize=(3,5))\n",
    "plt.errorbar(x=geneva_outcome_odds_filt[\"quant\"].tolist(),y=geneva_outcome_odds_filt[\"odds\"].tolist(),yerr=geneva_outcome_odds_filt[\"se\"].tolist(),fmt=\"o\")\n",
    "plt.xlabel(\"LASSO signature PRS quantile\")\n",
    "plt.ylabel(\"odds of melanoma\")\n",
    "plt.title(\"High-Density Melanoma Cohort\")\n",
    "plt.xticks([0,1])\n",
    "plt.xlim((-1, 2))\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.savefig(\"../plots/fig3/1vs9_melanoma.quantile.pdf\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4868e038-be48-4af6-a506-c6b8ec955ab7",
   "metadata": {},
   "source": [
    "## Melanoma based TCGA inflammation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "f7407215-e10c-45ae-910f-9f7900b40d29",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "124 PRS snps\n",
      "0 snps missing weight\n",
      "switching 108 genotypes\n"
     ]
    }
   ],
   "source": [
    "group1_sumstats=ukbb_annot[ukbb_annot[\"group\"]==\"group1\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "group1_sumstats=group1_sumstats[group1_sumstats[\"Outcome\"]==\"C43-C44\"]\n",
    "group1_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "mhcII_sumstats=ukbb_annot[ukbb_annot[\"file\"]==\"Sigs160-Wolf_MHC2-21978456\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "mhcII_sumstats=mhcII_sumstats[mhcII_sumstats[\"Outcome\"]==\"C43-C44\"]\n",
    "mhcII_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "group1_mhcII_sumstats=group1_sumstats.append(mhcII_sumstats)\n",
    "group1_mhcII_sumstats\n",
    "\n",
    "tcga_group1_prs,tcga_snp_table=make_ukbb_prs(\"../data/genotypes/all.time.raw\",group1_mhcII_sumstats,group1_mhcII_sumstats[\"ID\"].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "ff0bc895-4435-4091-9205-ce6f91269ed8",
   "metadata": {},
   "outputs": [],
   "source": [
    "canc = pd.read_csv(\"/cellar/controlled/dbgap-genetic/phs000178_TCGA/birdseed-processing/gtype.meta\", delimiter=\"\\t\", header=None)\n",
    "mp = canc.set_index(2)[0].to_dict()\n",
    "tcga_group1_prs[\"cancer\"]=tcga_group1_prs[\"IID\"].map(mp)\n",
    "tcga_group1_prs_SKCM=tcga_group1_prs[tcga_group1_prs[\"cancer\"]==\"SKCM\"]\n",
    "\n",
    "tcga_snp_table=tcga_snp_table[tcga_group1_prs_SKCM[\"IID\"]]\n",
    "tcga_snp_table=tcga_snp_table.transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "3935592c-fb0e-4554-b799-8e62efffad55",
   "metadata": {},
   "outputs": [],
   "source": [
    "#trim tcga_snp_table down to model feature space\n",
    "tcga_snp_table_test=tcga_snp_table[geneva_outcome_cov_table_x.columns.values]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "5c92a03c-a95e-4023-a8cb-3a62a4ee1e4a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "p = regressor_xgb.predict(tcga_snp_table_test)  # test the output by changing values\n",
    "tcga_snp_table_test[\"LASSO_burden\"]=p\n",
    "tcga_group1_prs_SKCM[\"LASSO_burden\"]=p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "46135bb1-92ab-423a-8567-6002d28deb44",
   "metadata": {},
   "outputs": [],
   "source": [
    "#get cibersortx data\n",
    "ciber_lm22=pd.read_csv(\"../data/plink-associations/phenotypes/TCGA.Kallisto.fullIDs.cibersort.relative.tsv\",delimiter=\"\\t\")\n",
    "ciber_lm22[\"id\"]=ciber_lm22[\"SampleID\"].str.replace(\".\",\"-\")\n",
    "ciber_lm22[\"code\"]=ciber_lm22[\"id\"].str.rsplit(\"-\",4).str[1]\n",
    "ciber_lm22[\"id\"]=ciber_lm22[\"id\"].str.rsplit(\"-\",4).str[0]\n",
    "ciber_lm22=ciber_lm22[ciber_lm22[\"code\"].str.contains(\"01\")]\n",
    "ciber_lm22.index=ciber_lm22[\"id\"]\n",
    "ciber_lm22=ciber_lm22[ciber_lm22.columns[2:-5].tolist()]\n",
    "ciber_lm22=ciber_lm22.groupby(ciber_lm22.index).mean()\n",
    "ciber_lm22=ciber_lm22.reset_index().rename(columns={\"id\":\"IID\"})\n",
    "tcga_group1_prs=pd.merge(tcga_group1_prs_SKCM,ciber_lm22,on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "4d9c2032-43c9-43b8-99e6-ca6399f8d8c7",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "missing T.cells.gamma.delta\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:2: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).\n",
      "  \n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:2: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).\n",
      "  \n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVAAAAFNCAYAAABWoDecAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAev0lEQVR4nO3de5RcZZ3u8e+TDhgCZLiMCgRMQALYqCBkAAUvIzAQVILgcAmKIhpRkJEZx4Oe4DXO8ijjUmaQGEYQ1ICCikEz4OhwVS4JF0E6XHKASCSAjMfcGoEkv/PHfls2RXVq95veXV3dz2etWl17v+/e+1fVXU/veykiMDOzgRvT7gLMzDqVA9TMLJMD1MwskwPUzCyTA9TMLJMD1MwskwPUzCyTA3QUkrSNpB9LWiNpqaQZG+h7pqTHJa2QdKGkl5TavitpuaSVkh6Q9IFS22RJIWl16XF2qX0rSRdLejI9Ptuw3MmSrpXUK+k+SYc0tH9U0sNp2YskHVRq+7KkR1PbUkn/u2HauZLul7Re0vsa2l4t6RpJT0l60UnSrd47SeMlfSNNv0LSDaW2/2x4P56VdE+pfW9JN6bplkn6dD+/k4vSe7trw/hDJN2RantU0rFp/F9L+pWk/5H0J0k3Szqw2bxtgCLCj1H2AC4Fvg9sARwErAD2bNLvMOAJYE9ga+A64Eul9j2Bl6TnewCPA/um4clAAGP7qeEi4HJgfOr7f4GTS+03A18FNgOOAf4EvDS17Q+sAfYFBHwY+APQldp3BzZPzycC9wJHl+Z9GnAwsAh4X0NduwOnANOLj8fA3jvgu8BlwEuBrr73o5/34Drg06XhHuCLabpXAsuBIxumOQi4Ib23u5bGdwNPAtOAscC2wCtT27j0usak9+so4I/9/W78GMBnqd0F+DHEv3DYHHgW2K007jvlYCyNnwf8S2n4YODxfua7e/rAH5uGWwXoU8DflIY/BdyYnu8GPANsWWq/ETg1PT8OuK3hNQWwfZPlTATuAT7RpO2mxgAtte3aGKCt3rv0HqwEJlT4PUwG1gE7l8b1At2l4cuBT5aGxwJ3Aq9tEqDzgC9UWO4Y4B1p+pe1+++x0x/ehB99dgPWRcQDpXG/oVibbLRnaiv3e7mkbftGpM3VXuA+igBd0DCPpWlz9CJJf93Qpobnry4t96GIWNVPjf8JdEnaX1IX8H7gLoo14L66zpK0GlhGEXzzmry+gWr13u0PLAU+lzbh75F0TD/zOoniH8bDpXFfA06StImk3YHXA78otZ8J3BARdzeZ3wEAaZnL0+6VbcodJN0N/BmYD/xHRDxZ5UVb/xygo88WFJudZSuALSv07Xv+l74R8ZE0/EbgRxRrjpDWMIFJFJvaWwLfK83rauAsSVumfXnvp9icr1LjKuCHFGuQzwCfAWZGWsVKdX0p9d+HYi2xcX45WtW1I8U/gRXADsDpwMWSXtVkXicB324Y91PgXcDTFP+QvhURCwEk7QR8CGi6XzQt+z0UuzumUOz6+Ldyh4h4LTABmEHx3tlGcoCOPqspPkRlEyhCqVXfvucv6BsR6yLiJooP8YfTuNURsSgi1kbEExRh8neS+uZxBkVQPAj8hGLf4rKKNX6AInD3BDYF3g38VNIODXVFRNyZlvO5Jq9voFrV9TTwHDA7Ip6NiOuBa4G/K0+QDnhtB1xRGrcNxT+Vz1Pss9wJOEzSR1KXrwGfj4j+/hE8DVwUEQ9ExGrgX4AjGjtFxJ8j4lKKf157VXvZ1h8H6OjzADBW0pTSuL0oDrQ0uje1lfs9ERH/08+8x1Ic/Gimb+1QABHxx4g4MSK2i4g9Kf4WbystdxdJ5bXico17AVelsFgfEVdT7D54Q0ZdA9HqvWu2ad3Me4EfpaDrswvF7oFL0j+dZRQHo/pC8GDgK+mMiL5dFTeXzgK4m+ff4yo2Scu0jdHunbB+DP2D4oN5KcW+wQPp/yj84RT7FbspjsL/N88fMHkZcDzFZm0XxRH7NcD01L4/zx/53ZbiyPW1pXm/Mo3vojhy/BQvPJp9C3AOxdrYO3nhUfj3UoTZLhSBfCjFAZg90vI+lOoVsB9FuJ5Rmvemab6/Aj6Yno9JbUrD3RSBNI50pkGr944ilJYAZ1OE9oEUa6d7lKbfLL2Wtza81xPS+BnpNWxHcSbCF0vv93alR1Ds99wstb8feDi9J+OBHwDfSW0HUBy93zQt/3+lunZo999ipz/aXoAfbfilwzbAlSnwfgfMSONfQbGZ+opS33+kOJVpJcWpR32nLb0UuD596FdSHOn+YGm6E9IHek0KsEuA7UrtxwKPpeC7CzisocbJFKf5PA3cDxxSahPFpu7vUhAsBt6T2sZQbAr/Mb2WByiO8Ks0/XUpgMqPt5SW29j2SKv3rtS+Zwq+NRSnJb2zof0EigNNavJ7eSuwkCKUHwcuAMb38zt8wVH4NO5zFKdz/YFiv+/WafybKQ52rUrvy/XAm9r9dzgSHkpvsJmZDZD3gZqZZXKAmpllcoCamWVygJqZZXKAmpllGtvuAgbT3kcd5VMKzKwWd115pRrHjagABbjryivbXYKZjTB7H3VU0/HehDczy+QANTPL5AA1M8vkADUzy+QANTPL5AA1M8vkADUzy+QANTPL5AA1M8s04q5EGs5mz55NT0/PkC1v6dKlAEyaNGnIltnd3c2sWbOGbHlm7eQAHcF6e3vbXYLZiOYAHUJDvWY2Y0bxhY3z5s0b0uWajRbeB2pmlqnWAJV0uKT7JS2RdFaT9j0k3SzpGUkfH8i0ZmbtVluASuoCzqP4zu9u4ARJ3Q3d/gicQfH93wOd1sysrepcA90PWBIRD0XEs8BlwPRyh4h4MiIWAs8NdFozs3arM0AnAo+WhpelcXVPa2Y2JOoM0Bfd/h6o+pUblaeVNFPSIkmLKldmZjYI6gzQZcBOpeEdgccGe9qImBsRUyNialaVZmaZ6gzQhcAUSTtL2hQ4Hpg/BNOamQ2J2k6kj4i1kk4HrgG6gAsj4l5Jp6b2OZK2AxYBE4D1kj4GdEfEymbT1lWrmVmOWq9EiogFwIKGcXNKzx+n2DyvNK2Z2XDiK5HMzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwy1Rqgkg6XdL+kJZLOatIuSeem9rsl7VNqO1PSvZJ+K+lSSePqrNXMbKBqC1BJXcB5wDSgGzhBUndDt2nAlPSYCZyfpp0InAFMjYhXA13A8XXVamaWo8410P2AJRHxUEQ8C1wGTG/oMx24JAq3AFtJ2j61jQU2kzQWGA88VmOtZmYDVmeATgQeLQ0vS+Na9omI3wPnAL8DlgMrIuLnNdZqZjZgdQaomoyLKn0kbU2xdrozsAOwuaR3N12INFPSIkmLNqpaM7MBqjNAlwE7lYZ35MWb4f31OQR4OCL+EBHPAT8C3tBsIRExNyKmRsTUQavczKyCOgN0ITBF0s6SNqU4CDS/oc984KR0NP4Aik315RSb7gdIGi9JwMHA4hprNTMbsLF1zTgi1ko6HbiG4ij6hRFxr6RTU/scYAFwBLAE6AVOTm23SroCuANYC9wJzK2rVjOzHLUFKEBELKAIyfK4OaXnAZzWz7SfAT5TZ31mZhvDVyKZmWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZar1RPpOMHv2bHp6etpdRi0WLy6ufp0xY0abK6lPd3c3s2bNancZNkqN+gDt6enhltvuYB0T2l3KoBvDswD86rYlba6kHl2sbHcJNsqN+gAFWMcEVo5perMnG8YmrP91u0uwUc77QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwyOUDNzDI5QM3MMjlAzcwybfBaeEmrgOivPSJG3h04zMwq2mCARsSWAJI+DzwOfAcQcCKwZe3VmZkNY1U34Q+LiG9ExKqIWBkR5wPH1FmYmdlwVzVA10k6UVKXpDGSTgTW1VmYmdlwVzVAZwDHAk+kx9+ncWZmo1alGypHxCPA9HpLMTPrLJXWQCXtJumXkn6bhl8ryV9EY2ajWtVN+AuATwLPAUTE3cDxdRVlZtYJqgbo+Ii4rWHc2sEuxsysk1QN0KckvZJ0Ur2kdwHLa6vKzKwDVP1WztOAucAekn4PPExxMr2Z2ahVNUAjIg6RtDkwJiJWSdq5zsLMzIa7qpvwPwSIiDURsSqNu6KekszMOkOrm4nsAewJ/JWko0tNE4BxdRZmZjbctdqE3x14O7AV8I7S+FXAB+sqysysE7S6G9NPgJ9Ien1E3DxENZmZdYSqB5HulHQaxeb8XzbdI+L9tVRlZtYBqh5E+g6wHXAYcD2wI8Vm/AZJOlzS/ZKWSDqrSbsknZva75a0T6ltK0lXSLpP0mJJr69Yq5nZkKgaoLtGxNnAmoi4GHgb8JoNTSCpCzgPmAZ0AydI6m7oNg2Ykh4zgfNLbV8Hro6IPYC9gMUVazUzGxJVN+GfSz//JOnVFHenn9ximv2AJRHxEICkyyju6NRT6jMduCQiArglrXVuD6wB3gS8DyAingWerVir2Ygwe/Zsenp6WnccJEuXLgVg0qRJQ7bM7u5uZs3q3PsSVV0DnStpa+BsYD5FCP6fFtNMBB4tDS9L46r02QX4A3CRpDsl/Uc6id/MatLb20tvb2+7y+goVddAL4qIdRT7P3epOI2ajGv8grr++owF9gE+GhG3Svo6cBZFgL9wBtJMis1/9pruW5bayDHUa2YzZhT3SJ83b96QLreTVV0DfVjSXEkHS2oWes0sA3YqDe8IPFaxzzJgWUTcmsZfQRGoLxIRcyNiakRMrViXmdmgqBqguwO/oLipyCOS/l3SQS2mWQhMkbSzpE0p7h86v6HPfOCkdDT+AGBFRCyPiMeBRyXtnvodzAv3nZqZtV3Vr/R4GvgB8IO0L/TrFJvzXRuYZq2k04FrUr8LI+JeSaem9jnAAuAIYAnQC5xcmsVHge+l8H2ooc3MrO2q7gNF0puB4yhOPVpI8SVzGxQRCyhCsjxuTul5UKzVNpv2LsCb5WY2bFUKUEkPA3dRrIX+c0SsqbUqM7MOUHUNdK+IWFlrJWZmHabV7ew+ERFfBr4oqfEUJCLijNoqMzMb5lqtgfZdPrmo7kLMzDpNq9vZXZWe9kbE5eU2SX9fW1VmZh2g6nmgn6w4zsxs1Gi1D3QaxXmaEyWdW2qagL8X3sxGuVb7QB+j2P95JHB7afwq4My6ijIz6wSt9oH+BviNpB9T3At0HfzlXp8vGYL6ard06VK6WMmE9b9udyk2QF2s/Mst2Mzaoeo+0J8Dm5WGN6O4Nt7MbNSqeiL9uIhY3TcQEaslja+ppiE1adIklj3xHCvHvKHdpdgATVj/6yG9+a9Zo6proGsavq9oX+DpekoyM+sMVddAPwZcLqnvfp7bU9xYxMxs1Kp6O7uFkvaguC+ogPsi4rkWk5mZjWiVb2dHEZ7dFN8L/zpJRMQl9ZRlZjb8Vb2d3WeAt1AE6AKKe4LeBDhAzWzUqnoQ6V0UX6vxeEScTPE97SPiPFAzs1xVA/TpiFgPrJU0AXiS6t/OaWY2IlXdB7pI0lbABRSXdK4GbqutKjOzDlD1KPxH0tM5kq4GJkTE3fWVZWY2/A3kKDwAEfFIDXWYmXWcqvtAX0TSHYNZiJlZp8kO0IjYp3UvM7ORq1KAStpc0pj0fDdJR0rapN7SzMyGt6proDcA4yRNBH4JnAx8u66izMw6QdUAVUT0AkcD/xYR76S4KsnMbNSqHKCSXg+cCPwsjRvwEXwzs5GkaoB+jOJbOH8cEfdK2gW4tr6yzMyGv6on0l8PXF8afgg4o66izMw6QauvNb4KiP7aI+LIQa/IzKxDtFoDPWdIqjAz60Ctvtb4+g21m5mNZq024e+h+Sa8gIiI19ZSlZlZB2i1Cf/2IanCzKwDtdqEX9r3XNIkYEpE/ELSZq2mNTMb6apeC/9B4Argm2nUjsCVdRVlZtYJqp5IfxpwILASICIeBF5WV1FmZp2gaoA+ExHP9g1IGssGzg81MxsNqgbo9ZI+BWwm6VDgcuCq+soyMxv+qgboWcAfgHuAD1F8N/ysVhNJOlzS/ZKWSDqrSbsknZva75a0T0N7l6Q7Jf20Yp1mZkOm6pH0zYALI+ICKIItjevtb4LU5zzgUGAZsFDS/IjoKXWbBkxJj/2B89PPPv8ALAYmVKzTzGzIVA3QXwKHUHydMRTh+XPgDRuYZj9gSbrxCJIuA6YD5QCdDlwSEQHcImkrSdtHxHJJOwJvA74I/GPVF2RWl9mzZ9PT09O6Y4davHgxADNmzGhzJfXp7u5m1qyWG8+VVQ3QcRHRF55ExGpJ41tMMxF4tDS8jBeuXfbXZyKwHPga8Algy4o1mtWqp6eHW+64lXVbtLuSeox5pvj5qwdubW8hNela3brPQFUN0DWS9omIOwAk7Qs83WIaNRnXeOS+aR9JbweejIjbJb1lgwuRZgIzAfaaPr1FSWYbZ90WsHLf9e0uwzJMuD37OzT7NZAbKl8u6UZJNwLfB05vMc0yYKfS8I7AYxX7HAgcKekR4DLgrZK+22whETE3IqZGxNSKr8XMbFBUvaHyQkl7ALtTrDXeFxHPtZhsITBF0s7A74HjgcadK/OB09P+0f2BFRGxnOLu958ESGugH4+Id1d7SWZmQ6Py9ewpMH/bNyxpu4h4fAP910o6HbgG6KI4in+vpFNT+xyK06GOAJZQHNE/OetVmJm1wcbcEORbFEfJ+xURCyhCsjxuTul5UFwmuqF5XAdcl1ukmVldNmav6tGDVoWZWQfaYIBKOruf8X9FsWluZjZqtVoDfaOkL5ZHSNoOuAF/rbGZjXKtAvRIYC9JXwWQNAW4CfhGRHyu7uLMzIazDQZoRPwZeCcwKZ1q9AvgnyPimxuazsxsNGj1pXJ916DfRnFZ5Y3Azn3jI+Kr9ZZnZjZ8tTqNqXwd+rlNxpmZjVqtvlTO+znNzPrRahP+0xtojoj4wiDXY2bWMVptwq9pMm5z4BRgW8ABamajVqtN+H/tey5pS4o7xJ9McYekf+1vOjOz0aDltfCStqG4I/yJwMXAPhHx/+ouzMxsuGu1D/QrFNe8zwVeU74rvZnZaNfqSqR/Anag+AbOxyStTI9VklbWX56Z2fDVah/o4N8D38xshHBAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllcoCamWVygJqZZXKAmpllqjVAJR0u6X5JSySd1aRdks5N7XdL2ieN30nStZIWS7pX0j/UWaeZWY7aAlRSF3AeMA3oBk6Q1N3QbRowJT1mAuen8WuBf4qIVwEHAKc1mdbMrK3qXAPdD1gSEQ9FxLPAZcD0hj7TgUuicAuwlaTtI2J5RNwBEBGrgMXAxBprNTMbsDoDdCLwaGl4GS8OwZZ9JE0GXgfc2mwhkmZKWiRp0UbWa2Y2IHUGqJqMi4H0kbQF8EPgYxGxstlCImJuREyNiKnZlZqZZagzQJcBO5WGdwQeq9pH0iYU4fm9iPhRjXWamWWpM0AXAlMk7SxpU+B4YH5Dn/nASelo/AHAiohYLknAt4DFEfHVGms0M8s2tq4ZR8RaSacD1wBdwIURca+kU1P7HGABcASwBOgFTk6THwi8B7hH0l1p3KciYkFd9ZqZDVRtAQqQAm9Bw7g5pecBnNZkuptovn/UzGzY8JVIZmaZHKBmZpkcoGZmmWrdB9opuljJhPW/bncZg24MawBYz+ZtrqQeXTQ9NdhsyIz6AO3uHrmX2C9evBiAV71q1zZXUp+R/Puz4W/UB+isWbPaXUJtZsyYAcC8efPaXInZyOR9oGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpkcoGZmmRygZmaZHKBmZpnGtrsAs06xdOlSulbBhNu93tGJulYVv8PB5L8EM7NMXgM1q2jSpEkse+ZxVu67vt2lWIYJt49h0qRJgzrPWtdAJR0u6X5JSySd1aRdks5N7XdL2qfqtGZm7VZbgErqAs4DpgHdwAmSuhu6TQOmpMdM4PwBTGtm1lZ1bsLvByyJiIcAJF0GTAd6Sn2mA5dERAC3SNpK0vbA5ArTdpzZs2fT0zN0L2Hx4sUAzJgxY8iW2d3dzaxZs4ZseUOta/XQHUQa0wtaNySLapvogvXjh2ZZXasHf551BuhE4NHS8DJg/wp9JlacFgBJMynWXtlr+vSNq3iEGT9+iP4yR4nu7qHdCFq6dCm9vb1DusyhNn78+EHfL7khg/07rDNA1WRcVOxTZdpiZMRcYC7A3kcd1bTPcDGS18xGA//+rFGdAboM2Kk0vCPwWMU+m1aY1sysrercmbMQmCJpZ0mbAscD8xv6zAdOSkfjDwBWRMTyitOambVVbWugEbFW0unANUAXcGFE3Cvp1NQ+B1gAHAEsAXqBkzc0bV21mpnlqPVE+ohYQBGS5XFzSs8DOK3qtGZmw4kv5TQzy+QANTPL5AA1M8vkADUzy+QANTPL5AA1M8vkADUzyzTibqi891FHtbsEMxslVJzLbiOVpEURMbXddVhn8N/LwHgT3swskwPUzCyTA3Tkm9vuAqyj+O9lALwP1Mwsk9dAzcwyOUDNOoykyZJ+20/b5yUd0mL6z0r6eIXlSNJ/S5qQhi+U9GTjsiVtI+m/JD2Yfm6dxr9G0rcrv7AO5AAdhiQN6vcH+oMwekTEpyPiF4M0uyOA30TEyjT8beDwJv3OAn4ZEVOAX6ZhIuIeYEdJrxikeoYdB+jo4A/CyNMl6QJJ90r6uaTNACR9W9K70vMjJN0n6SZJ50r6aWn6bknXSXpI0hn9LONE4Cd9AxFxA/DHJv2mAxen5xcD5atZrqL4Sp4RyQE6jKU1x69I+q2keyQdl8Z/Q9KR6fmPJV2Ynp8iaXaTWfmDMPJMAc6LiD2BPwHHlBsljQO+CUyLiIOAlzZMvwdwGLAf8BlJmzRZxoHA7RVqeXn6LjPSz5eV2hYBb6wwj47kAB3ejgb2BvYCDgG+Iml74Aae/6OcCPR92fVBwI1N5uMPwsjzcETclZ7fDkxuaN8DeCgiHk7Dlza0/ywinomIp4AngZc3WcY2EbFqI+t8EthhI+cxbDlAh7eDgEsjYl1EPAFcD/wNRUi+UVI30AM8kYL19cCvm8zHH4SR55nS83W8+L4W2sjpAdZKqpIRfX9/pJ9PltrGAU9XmEdHcoAOb00/BBHxe2Briv2YN1AE6rHA6n6C0h+E0ec+YBdJk9PwcRnzuB/YpUK/+cB70/P3UtpdBOwGND1jYCRwgA5vNwDHSeqS9FLgTcBtqe1m4GM8H6Afp/nmO/iDMOpExNPAR4CrJd0EPAGsGOBsfga8pW9A0qUUf3e7S1om6ZTU9CXgUEkPAoem4T5/m+YzMkWEH8PsQbEmCcUa6Fcoguse4LhSn1OAx9LzTYA1wNH9zO9s4AOl4UuB5cBzwDLglDR+W4qj7w+mn9uUpvl34B3tfm/8GNDf0Ralv6NvAGcOcPrtgf/aiOW/BLgFGNvu96Kuhy/lHAXS5vglEXFo5vQvodj/elBErB3U4qw2ks6k2JLYFLgT+GBE9A5wHscCV8fzp8ANZNopwMSIuG6g03YKB+go4Q+C2eBzgJqZZfJBJDOzTA5QM7NMDlAzs0wOUBvxJI24b5+14cEHkazjSTqb4oYpjwJPUVwb/naKy1oPpLhA4C7gHIpLFhcCH46IZyQ9AkyNiKckTQXOiYi3SPos8EqKew3sBHw5Ii4Y0hdmw57/M1tHS6F3DPA6ir/nO3j+xilbRcSb052JHgQOjogHJF0CfBj4WovZvxY4ANgcuFPSzyLisTpeh3Umb8JbpzsI+ElEPB3FfQCuKrV9P/3cneLuRQ+k4YspLottpW++TwHXUtz6zewvHKDW6TZ016E1Ffqs5fnPwbiGtsb9W97fZS/gALVOdxPwDknjJG0BvK1Jn/uAyZJ2TcPvobg0FeARYN/0/JiG6aan+W5LcVONhYNZuHU+B6h1tIhYSHGQ6DfAjyhu/Lyioc+fgZOByyXdA6wH5qTmzwFfl3QjxX0xy26juJPQLcAXvP/TGvkovHU8SVtExGpJ4ylu7zczIu7YyHl+luKuWOcMRo02MvkovI0Ec9Pd+ccBF29seJpV5TVQM7NM3gdqZpbJAWpmlskBamaWyQFqZpbJAWpmlskBamaW6f8DkH/H5mFWOJcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for x in tcga_group1_prs.columns[4:]:\n",
    "    f=plt.figure(figsize=(5,5))\n",
    "    pheno=x \n",
    "    tcga_group1_prs_cancer=tcga_group1_prs[tcga_group1_prs[\"cancer\"]==\"SKCM\"]\n",
    "    tcga_group1_prs_cancer=tcga_group1_prs_cancer[~tcga_group1_prs_cancer[x].isnull()]\n",
    "    \n",
    "    tcga_group1_prs_cancer[\"group\"]=np.where(tcga_group1_prs_cancer[\"LASSO_burden\"]>tcga_group1_prs_cancer[\"LASSO_burden\"].quantile(0.9),\"high\",\"medium\")\n",
    "    tcga_group1_prs_cancer[\"group\"]=np.where(tcga_group1_prs_cancer[\"LASSO_burden\"]<tcga_group1_prs_cancer[\"LASSO_burden\"].quantile(0.1),\"low\",tcga_group1_prs_cancer[\"group\"])\n",
    "\n",
    "    ax=sns.boxplot(x=\"group\",y=pheno,data=tcga_group1_prs_cancer,order=[\"low\",\"high\"])\n",
    "    labels=[item.get_text() for item in ax.get_xticklabels()]\n",
    "    labels=[x+\" (\"+str(len(tcga_group1_prs_cancer[tcga_group1_prs_cancer[\"group\"]==x]))+\")\" for x in labels]\n",
    "    ax.set_xticklabels(labels)\n",
    "    \n",
    "    try:\n",
    "        cat1 = tcga_group1_prs_cancer[(tcga_group1_prs_cancer[\"group\"]==\"low\")]\n",
    "        cat2 = tcga_group1_prs_cancer[(tcga_group1_prs_cancer[\"group\"]==\"high\")]\n",
    "        plt.title(str(mannwhitneyu(cat1[pheno], cat2[pheno])[1]))\n",
    "        plt.savefig(\"../plots/fig3/melanoma.PRS.\"+x+\".pdf\")\n",
    "    except:\n",
    "        print(\"missing {}\".format(pheno))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "40daba9e-d5fc-4b18-8caf-0444794a919f",
   "metadata": {},
   "outputs": [],
   "source": [
    "mac_plot=pd.melt(tcga_group1_prs_cancer,id_vars=[\"IID\",\"group\"],value_vars=[\"Macrophages.M1\",\"Macrophages.M2\"])\n",
    "mac_plot=mac_plot[mac_plot[\"group\"].isin([\"low\",\"high\"])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "96cbf361-ab0f-4bc5-b7ca-8ce258c275a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f=plt.figure(figsize=(10,5))\n",
    "sns.boxplot(x=\"variable\",hue=\"group\",y=\"value\",data=mac_plot)\n",
    "plt.ylabel('% Infiltration')\n",
    "plt.savefig(\"../plots/fig3/mac.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "2dfd77d6-a921-4233-acfa-cb88f9eb6ad8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "tcga_group1_prs_cancer=tcga_group1_prs_cancer.rename(columns={\"T.cells.regulatory..Tregs.\": \"CD4+ Tregs\", \"T.cells.CD8\": \"CD8+ T Cells\"})\n",
    "T_plot=pd.melt(tcga_group1_prs_cancer,id_vars=[\"IID\",\"group\"],value_vars=[\"CD8+ T Cells\",\"CD4+ Tregs\"])\n",
    "T_plot=T_plot[T_plot[\"group\"].isin([\"low\",\"high\"])]\n",
    "f=plt.figure(figsize=(10,5))\n",
    "sns.boxplot(x=\"variable\",hue=\"group\",y=\"value\",data=T_plot)\n",
    "plt.ylabel('% Infiltration')\n",
    "plt.savefig(\"../plots/fig3/Tcells.pdf\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f2bf682-a03a-4ebf-a675-330098bbcbd3",
   "metadata": {
    "tags": []
   },
   "source": [
    "### UKBB with combined MHC+group1 SNPs, protstate training and validation ONLY"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a08daab0-5757-4545-b512-56de0e8475fb",
   "metadata": {},
   "source": [
    "#### Prostate Genes ONLY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "24b6667f-f745-4d59-b0fc-108cbe18a986",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "57 PRS snps\n",
      "0 snps missing weight\n",
      "switching 53 genotypes\n"
     ]
    }
   ],
   "source": [
    "group1_sumstats=ukbb_annot[ukbb_annot[\"group\"]==\"group1\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "group1_sumstats=group1_sumstats[group1_sumstats[\"Outcome\"]==\"C60-C63\"]\n",
    "group1_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "mhcII_sumstats=ukbb_annot[ukbb_annot[\"file\"]==\"Sigs160-Wolf_MHC2-21978456\"][[\"proxy_variant\",\"ukbb_a1\",\"Var1_OR\",\"Outcome\"]]\n",
    "mhcII_sumstats=mhcII_sumstats[mhcII_sumstats[\"Outcome\"]==\"C60-C63\"]\n",
    "mhcII_sumstats.columns=[\"ID\",\"A1\",\"OR\",\"Outcome\"]\n",
    "\n",
    "group1_mhcII_sumstats=group1_sumstats.append(mhcII_sumstats)\n",
    "group1_mhcII_sumstats\n",
    "\n",
    "group1_mhc_prs,test_ukbb_res=make_ukbb_prs(\"../data/ukbb/time-snp/dosage-multi/ukbb_bialleleic.raw\",group1_mhcII_sumstats,group1_mhcII_sumstats[\"ID\"].tolist())\n",
    "icd=pd.read_csv(\"/cellar/users/mpagadal/data/ukbb/plato_input/immune_cancer_lev1_pheno.txt\",delim_whitespace=True)\n",
    "group1_mhc_prs=pd.merge(group1_mhc_prs,icd,on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "266beeec-462a-4844-b8c8-30aada541722",
   "metadata": {},
   "outputs": [],
   "source": [
    "#merge group1_mhc_prs with raw SNP data then format for model building\n",
    "group1_mhc_prs_short=group1_mhc_prs[[\"IID\",\"C43-C44\",\"C60-C63\",\"AGE\",\"SEX\",\"PC1\",\"PC2\",\"PC3\",\"PC4\",\"PC5\",\"PC6\",\"PC7\",\"PC8\",\"PC9\",\"PC10\"]]\n",
    "test_ukbb_res_t=test_ukbb_res.transpose()\n",
    "SNP_outcome_cov_table=pd.merge(test_ukbb_res_t,group1_mhc_prs,right_on=\"IID\",how=\"left\",left_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "c9c36518-9b87-4a62-965a-77b0fed8148e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(21701.0, 14960.0, 7291.0)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SNP_outcome_cov_table[\"Cancer\"]=SNP_outcome_cov_table[[\"C43-C44\",\"C60-C63\"]].max(axis=1)\n",
    "np.nansum(SNP_outcome_cov_table[\"Cancer\"]),np.nansum(SNP_outcome_cov_table[\"C43-C44\"]),np.nansum(SNP_outcome_cov_table[\"C60-C63\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "bb75c14f-6359-4eac-a188-8d262a064627",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Try joint gene score but slim down to melanoma only\n",
    "SNP_outcome_cov_table.columns.values\n",
    "SNP_outcome_cov_table[\"AnyCancer\"]=SNP_outcome_cov_table[['C73-C75', 'C81-C96', 'C40-C41',\n",
    "        'C50-C50',  'C00-C14', 'C45-C49',\n",
    "       'C30-C39', 'C97-C97', 'C76-C80', 'C60-C63',\n",
    "       'C15-C26', 'C51-C58', 'C43-C44', 'C69-C72', 'C64-C68']].max(axis=1)\n",
    "\n",
    "SNP_outcome_cov_table=SNP_outcome_cov_table[np.logical_or(SNP_outcome_cov_table[\"C60-C63\"]==1,SNP_outcome_cov_table[\"AnyCancer\"]==0)]\n",
    "SNP_outcome_cov_table[\"Cancer\"]=SNP_outcome_cov_table[[\"C60-C63\"]].max(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "6a6b9471-29a7-4c0c-8e19-078c95cfa3ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table=SNP_outcome_cov_table.reset_index()\n",
    "SNP_outcome_cov_table=SNP_outcome_cov_table[~SNP_outcome_cov_table[\"Cancer\"].isnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "f33802e9-29fd-4bd0-bacf-603e34d48a0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table = SNP_outcome_cov_table.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "df14a07f-daf5-4695-bf30-13ff3ac4c281",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(341376, 112)\n"
     ]
    }
   ],
   "source": [
    "# Separate majority and minority classes\n",
    "df_majority = SNP_outcome_cov_table[SNP_outcome_cov_table[\"Cancer\"]==0]\n",
    "df_minority = SNP_outcome_cov_table[SNP_outcome_cov_table[\"Cancer\"]==1]\n",
    " \n",
    "# Downsample majority class\n",
    "df_majority_downsampled = resample(df_majority, random_state=random_sate,\n",
    "                                 replace=False,    \n",
    "                                 n_samples=len(df_majority))\n",
    "print(df_majority_downsampled.shape)\n",
    "\n",
    "# Combine minority class with downsampled majority class\n",
    "SNP_outcome_cov_table_downsamp = pd.concat([df_majority_downsampled, df_minority])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "76f3a26c-3a9c-42b1-9a97-93c150d0bc90",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_downsamp.drop([\"IID\",\"Cancer\",\"C43-C44\",\"C60-C63\",\"AGE\",\"SEX\",\"PC1\",\"PC2\",\"PC3\",\"PC4\",\"PC5\",\"PC6\",\"PC7\",\"PC8\",\"PC9\",\"PC10\",'PRS',\n",
    "       'FID', 'D55-D59', 'D80-D89', 'E70-E90', 'I70-I79', 'K20-K31',\n",
    "       'K50-K52', 'K70-K77', 'K80-K87', 'K90-K93', 'L40-L45', 'M00-M03',\n",
    "       'M05-M14', 'M15-M19', 'M20-M25', 'M70-M79', 'M45-M49', 'M60-M63',\n",
    "       'R70-R79', 'C73-C75', 'C81-C96', 'D60-D64', 'C40-C41', 'D37-D48',\n",
    "       'D70-D77', 'C50-C50', 'D65-D69', 'C00-C14', 'C45-C49', 'D00-D09',\n",
    "       'C30-C39', 'C97-C97', 'D50-D53', 'C76-C80', 'C15-C26', 'C51-C58',\n",
    "       'D10-D36', 'C69-C72', 'C64-C68','AnyCancer'],axis=1)\n",
    "SNP_outcome_cov_table_training_x.index=SNP_outcome_cov_table_downsamp[\"IID\"]\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_training_x.drop([\"level_0\",\"index\"],axis=1)\n",
    "SNP_outcome_cov_table_training_y=SNP_outcome_cov_table_downsamp[\"Cancer\"]\n",
    "SNP_outcome_cov_table_training_y.index=SNP_outcome_cov_table_downsamp[\"IID\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f603fee6-bc0d-44f8-8885-c35601f8c06a",
   "metadata": {},
   "source": [
    "### Generating and preparing the validation datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "5c2359fa-efa3-4ae3-9234-e4fc78f284df",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "57 PRS snps\n",
      "0 snps missing weight\n",
      "switching 45 genotypes\n"
     ]
    }
   ],
   "source": [
    "ellipse_prs,ellipse_snps=make_ukbb_prs(\"../data/genotypes/ellipse/ellipse.raw\",group1_mhcII_sumstats,group1_mhcII_sumstats[\"ID\"].tolist())\n",
    "#add phenotyping\n",
    "pheno=pd.read_csv(\"/cellar/controlled/dbgap-genetic/phs001120.v1.p1_ellipse/phenotypes/pheno_affection_status.txt\",delimiter=\"\\t\",header=None)\n",
    "pheno[2]=pheno[2].map({1:0,2:1})\n",
    "\n",
    "mp_pheno=dict(zip(pheno[1],pheno[2]))\n",
    "\n",
    "ellipse_prs[\"pheno\"]=ellipse_prs[\"IID\"].map(mp_pheno)\n",
    "cov=pd.read_csv(\"/cellar/controlled/dbgap-genetic/phs001120.v1.p1_ellipse/covs/age-pcs-fh\",delimiter=\"\\t\")\n",
    "ellipse_prs=pd.merge(ellipse_prs,cov,on=\"IID\",how=\"left\")\n",
    "ellipse_prs[\"FH_PROS\"]=ellipse_prs[\"FH_PROS\"].replace(-9,0)\n",
    "\n",
    "ellipse_prs[\"FID\"]=ellipse_prs[\"IID\"]\n",
    "\n",
    "#Merge geneva_prs file with SNP inputs, then chop up into train/test\n",
    "ellipse_snps_t=ellipse_snps.transpose()\n",
    "ellipse_snps_t[\"IID\"]=ellipse_snps_t.index\n",
    "\n",
    "ellipse_snps_t = ellipse_snps_t.reset_index()\n",
    "ellipse_prs = ellipse_prs.reset_index()\n",
    "\n",
    "ellipse_snps_t[\"IID\"]=ellipse_snps_t['IID'].astype(str)\n",
    "ellipse_prs[\"IID\"]=ellipse_prs['IID'].astype(str)\n",
    "\n",
    "ellipse_outcome_cov_table=pd.merge(pd.DataFrame(ellipse_snps_t),pd.DataFrame(ellipse_prs),on=\"IID\",how=\"left\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "c65e6e34-ef9a-4ed6-b6d4-0749cab1b78b",
   "metadata": {},
   "outputs": [],
   "source": [
    "ellipse_outcome_cov_table=ellipse_outcome_cov_table[~ellipse_outcome_cov_table[\"pheno\"].isnull()]\n",
    "\n",
    "ellipse_outcome_cov_table_x=ellipse_outcome_cov_table.drop(['IID',\n",
    "       'PRS', 'pheno', 'age',\n",
    "       'FH_PROS', 'FID', 'PC1', 'PC2', 'PC3', 'PC4',\n",
    "       'PC5', 'PC6', 'PC7', 'PC8', 'PC9', 'PC10'],axis=1)\n",
    "\n",
    "ellipse_outcome_cov_table_x.index=ellipse_outcome_cov_table[\"IID\"]\n",
    "ellipse_outcome_cov_table_x=ellipse_outcome_cov_table_x.drop([\"index_x\",\"index_y\"],axis=1)\n",
    "\n",
    "ellipse_outcome_cov_table_y=ellipse_outcome_cov_table[\"pheno\"]\n",
    "ellipse_outcome_cov_table_y.index=ellipse_outcome_cov_table[\"IID\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "f8e0e9af-168a-4af0-b695-9f866fd574ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "SNP_outcome_cov_table_training_y=pd.to_numeric(SNP_outcome_cov_table_training_y, errors='coerce').astype(\"Int32\")\n",
    "ellipse_outcome_cov_table_y=pd.to_numeric(ellipse_outcome_cov_table_y, errors='coerce').astype(\"Int32\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "88cb0283-d5c3-407d-9c4f-76babf5960c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Minor lack of overlap in SNPs from training to testing dataset, so need to pare down training table to SNPs that exist in the ellipse cohort\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_training_x[ellipse_outcome_cov_table_x.columns.values]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "f8789863-ea1f-42cf-8c1d-ea6f52dd67da",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((348667, 50), (348667,), (91644, 50), (91644,))"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SNP_outcome_cov_table_training_x.shape,SNP_outcome_cov_table_training_y.shape,ellipse_outcome_cov_table_x.shape,ellipse_outcome_cov_table_y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "97befe41-7f0f-426a-9205-2f60eddea188",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.04647482, 0.04205098, 0.04489412, ..., 0.04191032, 0.03472614,\n",
       "       0.04742778])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "a3e62485-efb8-4966-be59-f780efa62169",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5382223014559917\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAJcCAYAAACixjPMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeZzNZf/H8deFtNCiaLmVUqQdRWlV2bO0EZKIbqnu9kX7vmhfLFF32mRJJEmIshSRKEuWRLITshvGXL8/PuM3c4uZM5zvNeeceT8fj3nM9ztzzvdzne7HY+63a3Xee0REREQksRTK7waIiIiIyD8ppImIiIgkIIU0ERERkQSkkCYiIiKSgBTSRERERBKQQpqIiIhIAlJIE5G95pxr4ZwbHtGz33fOPRPFs0VEEplCmojExDl3gXNunHNurXNutXPue+dcVQDv/cfe+9r53cadOee8c65cDr8v5pzb4Jwbsovf7fbzOueKOudecc4tynz/fOfcazu9v7VzbppzbpNzbplz7i3n3CExtPmJzHafvYuf98ztMzrn6jjnxjjn1jvnVjrnRjvnGuVWV0QSj0KaiOTKOXcQMBjoBBwKlAaeBNLys11x0Bj7DLWdc0ft+GEMn/dBoApwNnAgcAkwJdv77wFeAO4DDgaqAccCXzvniu6uMc45B7QEVgOt8vphnHONgX7Ah8DRwBHAY0DDvD5LRPKfQpqIxOJEAO99b+/9du/9Zu/9cO/9VPj/XqPvdrw4s3fnFufcb5k9Ok87505wzo13zq1zzn2yI6w45y7O7JF6yDn3l3PuD+dci901xDnXwDn3s3Pu78yerjP24nO1AroBU4HsNXP8vEBV4DPv/RJv/vDef5jZvoOwQHeb936o936b9/4P4BosqF2XQ3suBP4F3AE0yynQ7Swz4L0KPO29/6/3fq33PsN7P9p7/+9YnyMiiUMhTURiMQfY7pz7wDlXzzlXIob31AXOwnqR7gfexoLQMcBpQPNsrz0SKIn1WLUC3nbOVdj5gc65M4EewE3AYUB3YJBzbt+8fiDnXBngYuDjzK/rs/06t8/7A3B3ZhA9PTMg7XAesB8wIPsbvPcbgK+AWjk0qxXwBdA3875BHj5SBey/7ad5eI+IJDCFNBHJlfd+HXAB4IF3gJXOuUHOuSNyeNsL3vt13vsZwHRguPd+nvd+LRZWKu/0+ke992ne+9HAl1jP087+DXT33k/I7OH6ABuCrLYHH+t6YKr3/legN3Cqc65yjJ/3eWw4swUwCVjsnNsxPFkS+Mt7n76Lmkszf/8PzrkDgCZAL+/9Nixs5WXI87BsNUQkBSikiUhMvPczvfetvfdHYz1h/wJez+Ety7Ndb97FffFs92u89xuz3S/IfP7OjgXuyRzq/Ns59zfWe7Sr1+bmeqwHDe/9EmA02UJRTp83MyB28d6fDxwCPAv0cM6dDPwFlHTOFdlFzaMyf78rVwLpwI5FDB8D9ZxzpTLv04F9sr/BObfjfhuwKlsNEUkBCmkikmfe+1nA+1h4iYcSzrli2e7LAEt28bqFwLPe+0OyfR3gve+dl2LOufOA8sCDmSsvlwHnAM13Fa5y+ryZ89W6AGuAU4DxWO/eVTvVLAbUA0buplmtsOD6Z2Z7+mGhbMew8J/AcTu9pyywHVgMzMb++1ydw0cXkSSikCYiuXLOneScu8c5d3Tm/TFYePghjmWezNza4kJsLla/XbzmHaC9c+4cZ4o55+o75w7M4blFnXP7ZfsqjAWir7FQVSnz6zTgAKz3KsfP65y7M3PBw/7OuSKZQ50HAlMyh3OfBDo55+o65/Zxzh2X+XkWAR/t3EDnXGmgRubn3tGeitiQ6o7evaFABedcy8xnHgo8B3zqvU/33nvgbuBR59wNzrmDnHOFnG0l8nYu/+1FJAEppIlILNZjPU0TnHMbsbAyHbgnTs9fhvVELcGG+dpn9l79D+/9JGxeWufM188FWufy7BnY8OqOr5ux+W6dvPfLsn3NxwJUK3L/vJuBVzLb/RdwK3C1935eZjtfBB4CXgbWAROwXq4a3vtdbVvSEvg5cwXp/7cJeBM4wzl3mvd+BXAZtmhiRWZ71mZ+nh3/fT4FmgJtMv9bLgeeAT7P5b+RiCQgZ//4EhHJH865i4GemXO/REQkk3rSRERERBJQZCHNOdfDObfCOTd9N793zrk3nXNznXNTM/c/EhERERGi7Ul7H9vMcnfqYaurygPtgLcibIuIJCjv/SgNdYqI/FNkIc17PwY7f253Lgc+zDxS5QfgkOxn54mIiIgUZLvabDGU0thqpx0WZf7sH7tlO+faYb1tVLz88rOCtE5EREQkDn4eONDl/qp/ys+QtqsG73Kpqff+bezcPypdcYX/eeDAKNslIiIikmd//w233ALLen9DYbaz5JRa7FP+ij1+Xn6u7lyEHeeyw9HseodxERERkYQ2Zw5UqACreg/jq0L1+eKMR5gxLWOvnpmfIW0QcH3mKs9qwFrvvQ4GFhERkaThPbz4ogW0qisGM7hQI/Y94yT2G/klFNq7mBXZcKdzrjdwMXbQ8CLgcTIPB/bed8MOEb4M2zF8E3BDVG0RERERibfly+H002HlSmh76Ge8va4phSpVhGHD4NBD9/r5kYU0733zXH7vsaNURERERJLKb79Bu3YW0Nq1g25+KG7aWTB0KBx8cFxq5OfCAREREZGksmkT3Hgj9O5t93ffsoVXuuwH27vCli1QrFjcaulYKBEREZEYbNkCVataQDvvPFjRsQevDDsNFi+GwoXjGtBAIU1EREQkV99+C8cdB7/+Cg89BN9d141SD7SFE06AEiUiqamQJiIiIpKDL76AWrVssWa/fvDsEW/ibrkZ6teHzz+HAw6IpK5CmoiIiMhu3H8/NGoE//oXTJsGjTd/BHfcAVdeCQMGwH77RVZbIU1ERERkJ2vWwO23w0sv2Ty0iRPhsMOABg3g4Yehb18oWjTSNiikiYiIiGTTt69tc9apE1x0EXw31nPk1x/ZyoESJeCZZ2CffSJvh0KaiIiICPDWW3DWWdCsmd0//zyMHuUp+viDcP318N//Bm2P9kkTERGRAmvzZsteDz8M69dDkSJw7bXQrRscWNzD3XfD669D+/Z2enpACmkiIiJSIE2fDo0bw+zZNr2sUSPo39+CGhkZ8J/boGtXm5z2+uvgXND2abhTREREChTvoUcPOOMMC2ivvGLTzT7/PDOgASxaBH36wH335UtAA/WkiYiISAEybRrUqGFnbgKMGwfnnpvtBRkZFsjKlIGpU23vjXwIaKCeNBERESkAMjJsO40zzrCAdt55sHz5TgEtPR2uuw6eesruS5fOt4AGCmkiIiKS4qZNg3POsY1pS5aEb76B77+Hww/P9qJt22xZZ+/ekW5Qmxca7hQREZGUNW2a9Z4BtG1r22z8Y4uztDS45hoYNAhefRXuuit4O3dFIU1ERERS0s8/Q+XKdv3TT3Dmmbt4kffQpIkd0NmlS/BtNnKi4U4RERFJOS+/nBXQnnpqNwENbM7ZNdfA228nVEAD9aSJiIhIiund23bOABg4EC6/fBcv2rABpkyBCy+0xQIJSD1pIiIikhLS0qBpUzsxoHx5mD9/NwFt7VqoUwfq1cvaiyMBqSdNREREkt6mTXD66TBvnp0i8NZbtpLzH9assYA2ZYptVluqVPC2xkohTURERJJaRgZceqkFtObNoVev3bxw1SqoVQtmzLDznxo1CtrOvFJIExERkaSVkQFVqljHWK1aOQQ0sJPUf/3Vzn+qWzdYG/eU5qSJiIhIUpo3D44+2gJaxYowbFgub7jvPtuLIwkCGiikiYiISBKaMsWGOJcuhauvtvtdnuC0eDHUrGmJrlAhOPXU4G3dUwppIiIikjS8h1tvtX3PFiyAu++GTz/dTUBbsAAuuggmTrSDOpOM5qSJiIhI0rjhBvjgA+sUmzDB5qPt0rx5cMkltt3GiBFw9tlB2xkPCmkiIiKS8JYvh4cftoBWpoxlsMKFd/PiefOsB23zZjtNfbfHDSQ2DXeKiIhIQps0CU46Cd59Fxo2hN9+yyGggW2QVrkyfPtt0gY0UE+aiIiIJLCpU6F2bfj7bxg50hYL7Nbs2bbc86CD7MD0JKeeNBEREUlIixbZ1hpr1th5nDkGtClT4Pzz4eabg7UvagppIiIiknCGD4cTT7Trzp2hWbMcXjxxoiW4YsXg8ceDtC8EhTQRERFJKD/9BFddZfP+X3zRttzYrXHjbB+0EiVgzBg44YRg7Yya5qSJiIhIwhg40Dan9d4WDJx1Vg4vTk+H66+HI4+0VZxHHx2snSEopImIiEi+27oV7rwT3nrL7ocPzyWgARQpYudwHnooHHVU5G0MTcOdIiIikq8mTIBy5SygHXEETJ5sh6Xv1tCh8Mgj1t126qkpGdBAIU1ERETyUZ8+UK0aLFxoiwOWLbMtznbriy/g8svhyy9h06Zg7cwPGu4UERGRfNGzJ7Rsaedu/vILnH56Lm/o39+SXOXKMGyYreZMYepJExERkaC2bIF77rGAVqgQzJoVQ0Dr0weaNrUzOL/+2lZzpjj1pImIiEgw3kOLFjBggN3PnQtly8bwxkKFoHp1W/554IGRtjFRqCdNREREgsjIsHPPBwyAG26wHrVcA9qff9r3a66BESMKTEADhTQREREJ4Jdf7FD0776z05vefRf23TeXN731FpQvb28Cm7xWgCikiYiISKR+/x2aNLHrevXsYIBc89Ybb8Att0CdOlC1auRtTEQKaSIiIhKZNWtsD7TffoP77oMhQ2x6WY5efNF2tr36avj00xi63FKTQpqIiIhEYskSqFjRrl94wbJXrkaMgA4dbKuNPn2gaNFI25jItLpTRERE4m7lSjjlFFi7Frp3h3btYnxjjRq2gVqzZjaJrQBTT5qIiIjE1axZcOaZFtDefDOGgOY9PPsszJ5tk9VatCjwAQ0U0kRERCSOJk60g9EXLYJeveC223J5g/dw1112FmfPnkHamCwU0kRERCQuevWyfdDS0mDQIGjePJc3ZGTArbfaSs4774SnngrSzmShOWkiIiKy1z78EFq1suupU2M45mn7drjpJtswrUMHeP75ArcPWm7UkyYiIiJ7bONGC2c7Atr8+TEENICtW2HOHHj0UQW03VBPmoiIiOyRzZtt/tns2XYwwMSJcMghubxp2zYbDy1e3A5KL6B7oMVCPWkiIiKSZ+vXwwknWEBr2NC+5xrQtm61rTUuuwzS0xXQcqGQJiIiInkydCgcfzwsXQqnnWaLBHIdrUxLg8aN7XT1q6+GIhrMy41CmoiIiMTs88+hfn346y9bLDBtWgxv2rwZLr8cvvjCDk2/447I25kKFGNFREQkV7/9Zoek//JL1n25cjG+uV07GD7cVnK2aRNZG1ONQpqIiIjkaN48qFIF1q2DatXglVfyENDANqqtX9/mo0nMNNwpIiIiu/Xjj1CpkgW0Tp1g/Hg477wY3rh2rW1S6z1UqKCAtgcU0kRERGSXhg2znjPvoW9f+M9/YnzjmjVQqxbcd1+Mk9ZkVxTSRERE5B9eftl2yjjqKMtZ11wT4xv/+gsuvdQmr/XvD2ecEWk7U5nmpImIiMj/6N3bOsEAJk+Gww+P8Y0rVkCNGjB3ri0DrVs3sjYWBOpJExEREcDOO3/vPbj2WqhYEf74Iw8BDeDnn2HhQhg8WAEtDtSTJiIiIgA895wdpXnssTYf7YgjYnxjWpqdHlC7tiW7XI8ekFioJ01ERKSA+/13m3/25JNQogRMnZqHgPbHH3bswKef2r0CWtyoJ01ERKQAe/99uOGGrPvvv4eDDorxzb//bosE1q2z7jeJK/WkiYiIFEAZGbY4YEdAmzDBtto4+eQYHzB7Nlx0EWzcCN98A1WrRtbWgko9aSIiIgVQ+/bwzjtw5JEwejSceGIe3rx8OVSvbqnu22/h9NMja2dBppAmIiJSwHTqZAHt6KPhzz/BuTw+4PDD4fbb4cor89D1JnmlkCYiIlKAjBsHHTrY9eTJeQxokydD0aK2UOChhyJpn2TRnDQREZECYvBgaNAANm+2AwFKlcrDmydMsEUCbdvaMKdETiFNREQkxXkPL74IV10F++wDP/2Ux9Oavv/ezuI87DD45JM9GB+VPaHhThERkRT34ovwwAN2/e23cMopeXjzqFHW/Va6tK3iLF06iibKLiikiYiIpKj0dNtm4/XXbSrZ+vX2PU9ee832QBsxwk5bl2AU0kRERFLQpk1wwgmwbBkULw6rVuUxoHlvw5q9e9teaHmawCbxoDlpIiIiKWb5cmjRwgJa3brw9995DGiDBtk+aOvWwQEHKKDlE4U0ERGRFDJyJBxzDAwcCP/5D3z1FRQunIcHfPopXH21HZq+fXtk7ZTcKaSJiIikgKVL4c47oU4d2LYNhg2DN9/M40N69YJmzeDss+Hrr+20dck3mpMmIiKS5KZNy9pS44IL4JVXLGflySefQMuWcOGFtqFa8eJxb6fkjXrSREREktg779jZ5kWKwHPPwdixexDQAKpUgeuugyFDFNAShHrSREREktR118HHH9v1jz9azsqzb7+1RQLHHw8ffBDX9sneUU+aiIhIklm9Gs46KyugLV68hwHttdfsqKd3341r+yQ+FNJERESSyKxZULmynXVety5s3Qr/+tcePKhjR7j7bmjcGFq3jnczJQ403CkiIpIkJkyAatXsun9/O4szz7yHp5+Gxx+H5s3hww9tQpskHPWkiYiIJDjvbf7ZjoDWqdMeBjSAuXPh2WehVSv46CMFtASm/2VEREQS2Pz5ULMmzJtn9z/8AOecsxcPLF/euuTOOAMKqa8mkel/HRERkQTkvS0MKF/eAtqdd9qB6XsU0LyHu+6yoU2ASpUU0JKAetJEREQSzIgRUKtW1v1338H55+/hwzIy4JZboHt3uOeeuLRPwlCMFhERSSBdu0Lt2nZ9zTWwdu1eBLTt2+HGGy2gPfAAvPRS3Nop0VNPmoiISIJ4/XUblQQ76um00/biYRkZtjjg449tJefjj4NzcWmnhKGQJiIiks8WLLBes4kT7X7lSihZci8fWqgQlCsHzzwDDz+8122U8BTSRERE8tEnn0DTpln3ex3Qtm61JaEVKsATT+xt8yQfaU6aiIhIPunbNyugffihLcLcq4C2ZYttoHb++bBmTVzaKPlHPWkiIiKBbdkC994LXbrY/eLFe3i0U3abN8MVV8Dw4dCtG5QosdftlPylkCYiIhLQ99/DBRdk3f/+exwC2saN0LAhjBoFPXrADTfs5QMlEWi4U0REJJC5c22BAMCDD9r0seOPj8ODn38eRo+2MVMFtJShnjQREZEA+vWzYLZkCfz3v9C2bRwf/sgjcOml9iUpQz1pIiIiEXv3XetB+/13W80Zl4C2erX1mq1eDfvtp4CWghTSREREIuK9bVN24412P38+NGkShwevXGmhrHdvmDo1Dg+URKThThERkQh4b9tr9Otn93/+CcccE4cHL18ONWpYt9ygQXDxxXF4qCQi9aSJiIjE2bZt1nvWr59lqLS0OAW0JUvsgfPnw5dfZh3yKSlJPWkiIiJxNG8eVKtmI5JVq8LIkXZCU1xs3w5Fi8LQoXDhhXF6qCSqSHvSnHN1nXOznXNznXMP7OL3BzvnvnDO/eKcm+Gc07phERFJWps322lMK1dCx452FmdcAtqyZRbQjjkGpkxRQCsgIgtpzrnCQBegHnAK0Nw5d8pOL7sV+NV7XxG4GHjFOVc0qjaJiIhEZc0aqFgR0tOhVSvo0CFOD547F84+G+65x+7j1i0niS7K/6XPBuZ67+d577cCfYDLd3qNBw50zjmgOLAaSI+wTSIiInGXng6XXw6//Wbba7z/fpwePGsWVK8OmzZB69ZxeqgkiyhDWmlgYbb7RZk/y64zcDKwBJgG3OG9z9j5Qc65ds65Sc65SVE1VkREZE94D3ffDWPHwtVX20a1cTF9ui0SSE+3454qVYrTgyVZRBnS3C5+5ne6rwP8DPwLqAR0ds4d9I83ef+2976K975K/JspIiKyZxYvtjlonTrZdhuffBKnB6elQf36NrQ5ejScdlqcHizJJMrVnYuA7AuOj8Z6zLK7AejovffAXOfcfOAkYGKE7RIREdlrq1fDySfD+vVQs6btK+t21T2xJ/bd18ZMjz4aypeP00Ml2UTZk/YjUN45VzZzMUAzYNBOr/kTqAHgnDsCqADMi7BNIiIie23FCjjzTAtoPXvC11/HKaBNmADvvWfXl1yigFbARdaT5r1Pd879BxgGFAZ6eO9nOOfaZ/6+G/A08L5zbho2PNrBe/9XVG0SERHZWxMm2A4Y27bZNLEWLeL04O++g3r14KijoHlzO49TCrRIN7P13g8Bhuz0s27ZrpcA2i5ZRESSwuzZtlEtQI8edr55XHz7LTRoYPugjRypgCaAjoUSERGJyZgxULmyXfftG8eANnw4XHYZHHecreIsvfNGCFJQKaSJiIjkomtX265s82ZbwXnNNXF8+LRptkR01Cg48sg4PliSnUKaiIhIDmbNgltvtesffoAmTeL04PXr7fs999iDS5WK04MlVSikiYiI7EJ6Orzyim2zAfDLL3DOOXF6eL9+cPzxMHWq3WsOmuyCQpqIiMhOnn/etii79167/+orOOOMOD3844+hWTM46SQoWzZOD5VUpJAmIiKSzbhx8NhjsHy5HfGUng5168bp4e+9By1b2gS3oUPhwAPj9GBJRZFuwSEiIpIsvIeHH7ZeNIAFC6BMmTgWGDYM2rSBWrVg4EA44IA4PlxSkXrSRESkwNu0CRo2tIC2//52WHpcAxrYCQLPPQeDBimgSUzUkyYiIgXa+vVw3nkwfToULw5r19q55nHz3nu2UW2pUvDgg3F8sKQ69aSJiEiBNXUqVKliAa1DB1i3Ls4B7bnnbIjztdfi+FApKBTSRESkwPHe8lPFijBnDnTrBh07xumQ9B0FnnjCJrm1aAFPPRWnB0tBouFOEREpUP7806aHzZtn9zNn2m4YcZN9BULr1rZEtHDhOBaQgkI9aSIiUmAMHgw1a1pAK1vWtteIa0ADm+TWvz+0awfvvquAJntMPWkiIpLyvIfrroNevey+Vy9o3jzORTIy7Ougg2D8eChRIo7jp1IQKaSJiEhK+/xzuP56WxRw7LG2vcYxx8S5SEYGtG8PGzZAz55w6KFxLiAFkYY7RUQkZU2ebCcwrVsHV15piwTiHtC2b4e2beGdd2wMVb1nEicKaSIikpIeeQTOOgu2bIEff4QBA6Bo0TgXSU+3brr334cnn4RnnlFIk7jRcKeIiKQU7+G226BLF7sfMMD2QotEu3Y2we2557RRrcSdQpqIiKSMFSugaVMYNcru09Ii6D3LrnVr22ztjjsiLCIFlYY7RUQkJWzYYPPORo2CG2+0Yc5IAtqWLXZAOsBFFymgSWQU0kREJOllZNjxmOPGWW565x3Yd98ICm3aBI0awVVX2S64IhHScKeIiCS1LVugRg0LaBdemDXUGXcbNkDDhjB6NPToASefHFEhEaOeNBERSVrdusH++1tAq1vXAlokiyvXrbMCY8bYPmitW0dQROR/qSdNRESS0rXXQu/edt2kCfTtG+HuF199BRMmQJ8+VkwkAIU0ERFJKmvWwL//bcdj7rOPdXLtt19Exby35Ne0qe3jccIJERUS+ScNd4qISNIYNcqyUv/+dvbm5s0RBrSVK22S27hxdq+AJoGpJ01ERJLC9OlwySV2/dFHdmB6ZJYts9UI8+bZggGRfKCQJiIiCe/vv6FNG7uePh1OPTXCYosXw6WXwqJFMGRIVjIUCUzDnSIiktAefBCOPdbO32zePOKAtnw5VK8OS5fCsGEKaJKv1JMmIiIJ65VXoGNHux48GC67LOKChx0GF19sKxPOOSfiYiI5U0gTEZGE4z3cey+8+qrdR34G59y5UKwYHHUU/Pe/ERYSiZ2GO0VEJKGkp9vJS6++CiVLwsKFEQe0mTPtLKlmzSwdiiQI9aSJiEjC+PtvOOkkmxrWsKGdY14oyu6E6dNtFadz0LVrhLvhiuSdetJERCQheA9Vq1pAq1YtQECbMsXmnxUpYudxRroiQSTvFNJERCTfeQ+33mpTw9q0gfHjIw5o3sMdd8ABB1hAq1AhwmIie0bDnSIikq9WrbJTBP74w76//XaAos5Bv352ZMFxxwUoKJJ36kkTEZF8s3Klzdn/4w9o1crOMC9cOMKCY8daoW3b4IgjFNAkoSmkiYhIvpgxw47D/PVXeOYZeP/9iIc4R46EunVh4kRboSCS4BTSREQkuBkz4LzzYP16eOQRePjhiAsOGwYNGsDxx9sp7aVKRVxQZO8ppImISFBdu8Jpp8G6dXY05tNPR1zwyy9t47WTToJvv7VhTpEkoJAmIiJBZGTA3XfbKk6Ajz6CevUCFC5ZEi680IY7S5YMUFAkPrS6U0REgmjaFD791K4XLIAyZSIuOGuW9Z6dcw58/bU2qpWko540ERGJ1Jw5UKdOVkDbsCFAQOvZ0zan7dPH7hXQJAkppImISGQWL7ZN/YcPt4UCGzfaOeaR6tEDrr/eCjdsGHExkegopImISCQ2b7Y90JYutQ1qv//eNviPVPfu0LYt1KoFgwcHSIQi0VFIExGRuPMeataEefPgqafg3/8OUPTXX+Hmm6F+ffj8c9h//wBFRaKjhQMiIhJ3Dz4I48bBGWfAo48GKnrKKdZ7VrMmFC0aqKhIdNSTJiIicdWnD7zwApx/Pvz8c4CCL75o+58BXHaZApqkDIU0ERGJi61b4T//gebN4eCDbTVnpIsqvYfHH4cOHbJWcYqkEIU0ERGJi6eegi5doEIF23bjyCMjLOa9jak+9RS0aWPHGIikGM1JExGRvbJ1K1x3HfTrZwFt5swAPWj33AOvvQbt21syjPRkdpH8oZAmIiJ7zHvbimz4cLufODHAvrHew19/we23w+uva6NaSVkKaSIiskcmTYKrroKFC+0+IyPivJSRAatWQalS8N571numgCYpTP3DIiKSZ7162ZGYCxdCo0awfXvEeWn7dpt7dn9nECwAACAASURBVO65sG4dFC6sgCYpTz1pIiKSJ598AjfcYB1bU6fC6adHXDA93Y556t3bFgocdFDEBUUSg0KaiIjEbOBAaNrUrmfOhJNOirjg1q1w7bXQvz907GjbbYgUEAppIiISk/nzLS8BjB8fIKABPPKIBbRXX4W77gpQUCRxKKSJiEiuRo2CSy6x6yFDoFq1QIXvvx8qVcpKhyIFiBYOiIhIjj7/PCugvfce1KsXccFNm+CJJyAtDUqWVECTAkshTUREdmnOHGjcGK64wu6HDYPWrSMuumGDnb/59NMwdmzExUQSm4Y7RUTkH376CapUsetWreCBBwLMQVu71gLahAnQsyfUrBlxQZHEpp40ERH5H926ZQW0116D998PENDWrIFatezIgr597ZR2kQJOPWkiIvL/Pv4Ybr4ZDjwQvvwSLrwwUOFFi+DPP20lZ6NGgYqKJDaFNBERAeDxx22vWLBNao87LkDRjRuhWDHbEff33+1aRAANd4qIFHjbtkH79lkBbdasQAFt6VKoWhVefNHuFdBE/od60kRECrDx4+HKK2H5crtfuhSOPDJA4UWL4NJLYckSOwRURP5BPWkiIgXU+PFw3nkW0K65xnrUggS0BQugenVYtsz29ahePUBRkeSjnjQRkQLo66+z5ucPGgQNGwYqvHmz7Yy7ejWMGAFnnx2osEjyUUgTESlg1q2D2rXt+ssvbWuyYPbfH558Ek49Fc48M2BhkeSjkCYiUoCsWgV16tj1G28EDGgzZ9o8tFq1oGXLQEVFkptCmohIAbFsGZQpY3PPbroJbr89UOFp06BGDShe3JaOFi0aqLBIctPCARGRAmDECDjqKAtoHTrYqQJBTJlic9CKFoWhQxXQRPJAIU1EJMX98IONMgI89hh07Bio8MSJts1GsWIwejSceGKgwiKpQcOdIiIpbOFCOPdcux450jJTMH36QIkS8O23cOyxAQuLpAb1pImIpKiVK7MWUL77bsCAtn27fX/5ZZgwQQFNZA8ppImIpKAxY+Dww+Gvv+zkpTZtAhUeOdLO4VywAAoVglKlAhUWST0KaSIiKeann7I28f/wQ5uTFsTQodCgARQpYvuhiche0Zw0EZEUMn9+1kkC48ZlzUeL3BdfQOPGcMopdpxByZKBCoukLvWkiYikCO8toC1ZAp99FjCgjRgBV10FFSvCN98ooInEiUKaiEgKGDvW5qBNnw6vvgpXXBGweJUq0Lat9aCVKBGwsEhqU0gTEUlyjz0GF11kiwSqVYO77gpUeOhQOzD9kENsd9yDDw5UWKRgUEgTEUli/fvD00+DczBvHowfH6jwu+/awZ/PPx+ooEjBo4UDIiJJasgQaNXKrpcvD7jbRdeucOutULcuPPhgoKIiBY960kREktCSJdCwIWzcaMdjBgtor79uAa1hQxg4UFttiERIIU1EJMnMmQMnnAAZGbaKs1KlQIVXrYJnn4Wrr4ZPP4V99w1UWKRg0nCniEgS+eQTaN7cAtqbbwZcxek9HHaY7Yxbpgzss0+gwiIFl0KaiEgS8N7mn330kd2/9Ra0bx+o8GOP2fXTT1sXnogEoeFOEZEE5z00bZoV0GbMCBjQOnSAZ56BZcvsXkSCUUgTEUlw/ftDv352bvmWLXbyUuS8tw3XXnoJbrkFune3fT5EJBiFNBGRBDZqFDRpYtcTJgScq3/77fDGG3DnndC5MxTS/12IhKY5aSIiCWpHRtpxHXS3i3PPheLF4bnn1IMmkk8U0kREElCHDvDii3Y9cSJUrRqgaHo6/PyzncV57bUBCopITtR/LSKSYD7/HF5+2a5/+SVQQNu2DVq2hPPPh99/D1BQRHKjnjQRkQTSvz80bmwjjTNm2JZkkdu61TZfGzDAuu+0zYZIQlBIExFJEH/9ZQENYPr0QAEtLc1WJnzxhR35dMcdAYqKSCwU0kREEsCKFVC5sl137AjHHhuo8AcfWEDr2hVuvjlQURGJhUKaiEg+mzsX6tWzQ9MffNAWDQTz73/bxmsXXBCwqIjEQgsHRETy0YQJtjBg/nw7l/O55wIUXb/e5qDNnWvbayigiSQkhTQRkXywbBmcdx5UqwZ//w2ffZa1aW2k1q6FOnXsCIOpUwMUFJE9peFOEZHA1q2DihVtHlrRovDNN7bzReTWrLGANmUK9O0LV10VoKiI7KlIe9Kcc3Wdc7Odc3Odcw/s5jUXO+d+ds7NcM6NjrI9IiL5bcMGmwK2YgXUqGGLK4MEtFWr4NJLbeO1AQPg6qsDFBWRvRFZT5pzrjDQBagFLAJ+dM4N8t7/mu01hwBdgbre+z+dc4dH1R4RkURw//2weDEceCCMGBGwcNGicOihtlNu3boBC4vInopyuPNsYK73fh6Ac64PcDnwa7bXXAsM8N7/CeC9XxFhe0RE8tWPP8Jbb1lWWrUqUNFly2xn3B2pUOdwiiSNKIc7SwMLs90vyvxZdicCJZxzo5xzPznnrt/Vg5xz7Zxzk5xzkyJqq4hIpNatg+sz/8J9+WWgoosWwUUXZZ3DqYAmklSiDGm7+mvgd7ovApwF1AfqAI865078x5u8f9t7X8V7XyX+zRQRidaIEVC2LMyaBV262IrOyP3xhwW05ctt8zURSTpRDncuAo7Jdn80sGQXr/nLe78R2OicGwNUBOZE2C4RkWCGD7cFlQA9e0KLFgGK/v67LRJYt84SYpAT2kUk3qLsSfsRKO+cK+ucKwo0Awbt9JrPgQudc0WccwcA5wAzI2yTiEgwf/4JzZrZ9bBhgQKa91Z040bb20MBTSRpRdaT5r1Pd879BxgGFAZ6eO9nOOfaZ/6+m/d+pnNuKDAVyAD+672fHlWbRERCyciABg1sa7LPPoPatQMVds7O49y+HU4/PVBREYlCpJvZeu+HAEN2+lm3ne5fAl6Ksh0iIiGtWAHnngvz5sEtt8AVVwQoOnWqpcHHHrON2EQk6elYKBGROOrRA44/3gLadddB584Bik6eDJdcAv/9L/z1V4CCIhKCQpqISJx07gxt29p0sHbt4KOPAux6MWGCLRI48EAYMwZKlYq4oIiEorM7RUTiYORIuO02OOAAWLAASpYMUPT776FePQtm33wDxx4boKiIhKKeNBGRvbBtmy0QqFnT7n/4IVBAA9sD7ZhjrAdNAU0k5SikiYjsocmT7UjML7+0rLR0aaAFlTvmnV11Ffz8M5Te+TAXEUkFMYc051yxKBsiIpJMxo2D6tXt+qabbIjzyCMDFB4yxI4vGDbM7vfZJ0BREckPuYY059x5zrlfydxk1jlX0TnXNfKWiYgkqIkT4fzzYcMGOzS9W7dAx2J+/rnt53HiiVBFp+SJpLpYetJew87VXAXgvf8FuCjKRomIJKJt22z/s3POsfuOHQNmpU8/hcaN4cwzbZXCYYcFKiwi+SWm1Z3e+4Xuf/+ZuD2a5oiIJKbFi60Ta9IkqFAB+vaFihUDFf/5ZzvqqVo1G+486KBAhUUkP8XSk7bQOXce4J1zRZ1z96LzNUWkAFm/Hi66yALaI4/ArFkBAxpYsU6dYOhQBTSRAiSWkNYeuBUoDSwCKgG3RNkoEZFE8cknNk9/3jzo0gWefjpg8Q8+gJkzbcLbzTdD8eIBi4tIfoslpFXw3rfw3h/hvT/ce38dcHLUDRMRyW8ffghNm8KqVfDKK3YOZzBdukDr1vDiiwGLikgiiWVOWifgzBh+JiKSMsaOhVat7Hr6dDj11IDFX3sN7r4bLr/clo6KSIG025DmnDsXOA8o5Zy7O9uvDgIKR90wEZH88sUX0KiRXf/2G5QrF7B4x47w4IO2krNXL+2DJlKA5dSTVhQonvmaA7P9fB3QOMpGiYjklylTsgLaTz8FDmjp6bZJbfPmNtZaRMcrixRku/0L4L0fDYx2zr3vvV8QsE0iIvnihx+gYUO77t3btiQLwntIS4P99rMzpvbdFwprwEKkoIvln2mbnHMvAacC++34off+0shaJSISWKdOcPvtdv3yy7YtWRDew/332zlTX38NBxwQqLCIJLpYVnd+DMwCygJPAn8AP0bYJhGRoHr2zApoQ4fCPfcEKuw93HmnpcJKlawnTUQkUywh7TDv/bvANu/9aO99G6BaxO0SEYmc9xbOWra0+6+/hjp1AhXPyLC9z958E+66Czp3hkKx/EkWkYIiluHObZnflzrn6gNLgKOja5KISPQ2bYLTT7dNas86C7p3t+/BPPSQFX3gAXjuuUAntItIMoklpD3jnDsYuAfbH+0g4M5IWyUiEqFZs6BBAwto5crBxIn50Il1441QsqSNrSqgicgu5PpnyXs/2Hu/1ns/3Xt/iff+LGB1gLaJiMTdr79aQPv9d3j0UdsHLVhA27YN3nvPxlnLlYN771VAE5Hdymkz28LANdiZnUO999Odcw2Ah4D9gcphmigiEh9Tp2YdjP744/DEEwGLb91qS0Y/+wyOOw4uuSRgcRFJRjkNd74LHANMBN50zi0AzgUe8N4PDNE4EZF4uv9++965M9x6a8DCW7bYCQJffglvvKGAJiIxySmkVQHO8N5nOOf2A/4Cynnvl4VpmohI/Hz0kW3mf/vtgQPapk1w5ZUwfLidw3nTTQGLi0gyyymkbfXeZwB477c45+YooIlIMvrxR7j+eihWDJ59NnDxyZNhzBjo0QNuuCFwcRFJZjmFtJOcc1Mzrx1wQua9A7z3/ozIWycishe8h9mzrSMLYNo0KF48UPGMDFuRcMEFtkrhX/8KVFhEUkVOIe3kYK0QEYmzLVvg4othwgS7f+stKFs2UPG//7YlpLfeaoelK6CJyB7I6YB1HaouIklp6VK44grb/6xZM2jTBmrVClR89WqoXduWkt57b6CiIpKKYtnMVkQkafzyC1SvDmvXQuXK0KtXwK3IVq60NDhrlm21Ub9+oMIikop0UJyIpIxVq2yIc+1aeP11m7MfLKBt2GBba8yeDYMGKaCJyF6LqSfNObc/UMZ7Pzvi9oiI7LFmzWw62H33wR13BC5erBg0aQIXXgiXXhq4uIikolx70pxzDYGfgaGZ95Wcc4OibpiISF60aAEjRsCRR8KLLwYsvHChLRt1zo4xUEATkTiJpSftCeBsYBSA9/5n59xxkbVIRCQPVq+GatXsDE6AGTMCFv/jDxviLFIEZs607yIicRLLnLR07/3ayFsiIpJH8+bBYYdZQKteHTZuhEMPDVR87ly46CKbANe7twKaiMRdLCFtunPuWqCwc668c64TMC7idomI5Gj6dDj1VLuuWxdGjYIDDghUfPZsS4WbNsE330CVKoEKi0hBEktIuw04FUgDegFrgTujbJSISE46doTTT7cNa199Fb76KnADnn0W0tMtGVaqFLi4iBQUsfTPV/DePww8HHVjRERysn27beDfr5/dDxtm+8YG1707LFkCJ5yQD8VFpKCIpSftVefcLOfc0865UyNvkYjIbjz9tAW0okVtUWXQgDZpEtSpY3PQ9t9fAU1EIpdrSPPeXwJcDKwE3nbOTXPOPRJ1w0REsuvVC558EkqUsGHOo48OWPyHH6BGDZgzxzZiExEJIKYTB7z3y7z3bwLtsT3THou0VSIimdLS4NprbR80gHHjAp4iAPDdd3bUU6lSMHo0HHtswOIiUpDlOifNOXcy0BRoDKwC+gD3RNwuEREArroKhgyx62XL4IgjAhYfO9aWjh5zDIwcCaVLBywuIgVdLD1p7wFrgNre++re+7e89ysibpeIFHDew/vvW0A7/3zIyAgc0MDCWfXqtopTAU1EAsu1J817Xy1EQ0REdvAeGjeGAQPsfuDAwEOcU6ZAxYpw3HFZ3XgiIoHttifNOfdJ5vdpzrmp2b6mOeemhmuiiBQkO3rMBgyA44+H33+HkiUDNmDgQDjnHHjppYBFRUT+KaeetDsyvzcI0RARkW3bbHRx5Uqbp//rr7DvvgEb0K+frVI46yy46aaAhUVE/mm3PWne+6WZl7d47xdk/wJuCdM8ESko/v4bTjkFxo+HM8+EFSsCB7SPP4Zmzey09uHD4ZBDAhYXEfmnWBYO1NrFz+rFuyEiUnD98YftfzZ3rt1PmhS4AcuXQ7t21o331Vdw0EGBGyAi8k85zUm72Tk3Daiw05y0+YDmpInIXvMe3nsv66D0hx+2OWlBFwmATYIbMQIGD4bixQMXFxHZtZzmpPUCvgKeBx7I9vP13vvVkbZKRFJeWho0bQqff273n30GV1wRuBGdO0OxYnDDDXDuuYGLi4jkLKfhTu+9/wO4FVif7Qvn3KHRN01EUlnLlhbQTj4ZNm7Mh4D2yitw223We+Z94OIiIrnLrSetAfAT4IHsAxAeOD7CdolICuve3RZS1qkDQ4fmQwOee87GVps0sQUDwcdXRURyt9uQ5r1vkPm9bLjmiEgqW78e2re3w9LBglpwTzxhJ7W3aGFHGhTJdU9vEZF8kevqTufc+c65YpnX1znnXnXOlYm+aSKSSjZsgJo1swLaokVw4IH50JAiRaB1a/jgAwU0EUlosWzB8RawyTlXEbgfWAB8FGmrRCSlbNoE5crBxIm2V+y2bYGPwvQe/vzTrh95BHr0gMKFAzZARCTvYglp6d57D1wOvOG9fwPIj3//ikgSSkuD006zrciuusqmgAXtwMrIgNtvh0qVsoKa5qCJSBKIJaStd849CLQEvnTOFQb2ibZZIpIKMjLgsstg/nzbwL9//3xoQPv2ttVGmzZwzDGBGyAisudiCWlNgTSgjfd+GVAa0MnDIpKjDRtsiPObb6BsWVizJnADtm+3YPbOO/DQQ3ZgunrQRCSJ5BrSMoPZx8DBzrkGwBbv/YeRt0xEktb8+XZA+vz5UKYMzJmTD43o3NkWBzz5JDzzjAKaiCSdWFZ3XgNMBJoA1wATnHONo26YiCSniROtB23LFrjoIliwIJ8WUbZvD336wGOPKaCJSFKKZbjzYaCq976V9/564Gzg0WibJSLJaN06OOccmwr21FMwenTgBqSlwX33wapVsO++du6UiEiSiiWkFfLer8h2vyrG94lIAeI91K5t1y+8AI+G/qfcli22fPTll+HrrwMXFxGJv1gGIYY654YBvTPvmwJDomuSiCSjjz+GCROgXj24//7AxTdtssM/R4ywM6eaNQvcABGR+Ms1pHnv73POXQVcgJ3f+bb3/rPIWyYiSWPmTDswvXjxfDjqacMGaNjQxlZ79LDTBEREUsBuQ5pzrjzwMnACMA2413u/OFTDRCQ5DB5sGQnsKMxixQI3YP16WLoUeva04wxERFJETj1pPYAPgTFAQ6ATcFWIRolIcnj3XbjxRrvu2ROuvjpg8XXrLBEedRRMnQpFiwYsLiISvZxC2oHe+3cyr2c75yaHaJCIJIfOneG222x3i99+gxNOCFh89WpbpVC5sm1Wq4AmIikop5C2n3OuMjYPDWD/7Pfee4U2kQLq1VfhnnvsetEi+Ne/AhZfuRJq1oTZs22fDxGRFJVTSFsKvJrtflm2ew9cGlWjRCRxdeqUFdAWLgwc0JYtgxo17CiDL76AWrUCFhcRCWu3Ic17f0nIhohI4hs8GG6/3a5XroSSJQMWz8iA+vXtCIMhQ+DiiwMWFxEJLz8OaxGRJDRgQNbCgJkzAwc0gEKF7JD0okXhggsCFxcRCU8nB4hIrnr0yDphado0OOmkgMXnz7eD0gEuvVQBTUQKDPWkiUiOxoyBtm3tevx4OO20gMV/+82C2ebNthnboYcGLC4ikr9y7Ulz5jrn3GOZ92Wcc2dH3zQRyW+rV9sxT2Dz0apVC1h85kyoXt3O5Bw5UgFNRAqcWIY7uwLnAs0z79cDXSJrkYgkBO/tCMxNm+w4zPr1AxafPt0WBmRkwKhRULFiwOIiIokhluHOc7z3ZzrnpgB479c457RzpEgK27YNjj4aVqywk5batQvcgNGjoUgR+OYbqFAhcHERkcQQS0/aNudcYWxvNJxzpYCMSFslIvlm/nwbZVyxwg5M//DDgMXT0uz7rbfCjBkKaCJSoMUS0t4EPgMOd849C3wHPBdpq0QkX0ydaiOL48dDhw52dnnhwoGKjx8P5cvDjz/a/SGHBCosIpKYch3u9N5/7Jz7CaiBHQl1hfd+ZuQtE5Ggxo6Fiy6y68GDA89BGzPGCh55pH2JiEjuIc05VwbYBHyR/Wfe+z+jbJiIhLNihS0SAOjWLXBAGzkSGjWCMmXsOug5UyIiiSuWhQNfYvPRHLAfUBaYDZwaYbtEJJDeveGGG2w62AcfwPXXByz+00/QoAGUKwcjRsARRwQsLiKS2GIZ7jw9+71z7kzgpshaJCLBjB1rqzcB3norcEADOOMMOwz0vvvy4ZwpEZHElucTB7z3k51zVaNojIiEM2aMreIsUsS2JQu6kPKrr+DMM63n7IUXAhYWEUkescxJuzvbbSHgTGBlZC0SkcilpcFll9l1586BA1rfvtCihXXb9egRsLCISHKJpSftwGzX6dgctf7RNEdEQrj9dti40bYjuynk5IWPPoLWreH88+GNNwIWFhFJPjmGtMxNbIt77+8L1B4Ridj338Pbb0Pp0taLFkyPHnDjjXDJJTBoEBQrFrC4iEjy2e1mts65It777djwpogkOe9h2DC44AK7HzQoYPG0NHj5Zahd2zZhU0ATEclVTj1pE7GA9rNzbhDQD9i445fe+wERt01E4qh7d7j5ZihUyALamaH++eU97LsvfPstHHww7LdfoMIiIsktljlphwKrgEvJ2i/NAwppIknizTfhjjvsesUKOOywQIVfesn2QuvZU3ugiYjkUU4h7fDMlZ3TyQpnO/hIWyUicfPQQ/D883b91VcBA9qzz8Ijj0DTptabJiIieZJTSCsMFOd/w9kO+osrkgQ6dswKaLNnw4knBijqPTzxBDz1FLRsaQsGiuR5S0YRkQIvp7+cS733TwVriYjEVY8e8OCDdr1xIxxwQKDCTz1lX23a2DLSwoUDFRYRSS05hbRd9aCJSBJ44gl48km7XrYsYEADqFkT1q611ZyFdruAXEREcpHTX9AawVohInHzwANZAW369EDz9TMyYORIuz7/fHj1VQU0EZG9tNu/ot771SEbIiJ7r2vXrKMwN2yAU08NUDQjw44tqFnTdsoVEZG40GxekRSwerXlpE8/tfutW2GffQIU3r4d2raFDz6Ahx+G884LUFREpGDQeIRIkvvrLyhb1gLaFVfAggWBAlp6uq3e/OADWyjwzDPgNJVVRCRe1JMmksTS06FOHVi3Di67DD77LGDxESOgd2/b56NDh4CFRUQKBoU0kSSVlmY9Z5Mnwy23QJcugRtQt64Vr1w5cGERkYIh0uFO51xd59xs59xc59wDObyuqnNuu3OucZTtEUkVCxbAscfC0KFQrRp07hyo8ObN0KQJjB1r9wpoIiKRiSykOecKA12AesApQHPn3Cm7ed0LwLCo2iKSSoYPhwoVYPly60kbPz7QVLBNm6BRI+jfH377LUBBEZGCLcqetLOBud77ed77rUAf4PJdvO42oD+wIsK2iKSECRNsDlpamm21EWwO2oYNNuntm2/gvffsNAEREYlUlCGtNLAw2/2izJ/9P+dcaeBKoFtOD3LOtXPOTXLOTYp7K0WSxMyZUK+eXQ8dCvffH6jwhg02/+y776BnT2jVKlBhEZGCLcqQFsvB7K8DHbz323N6kPf+be99Fe99lbi1TiSJzJwJVavCmjXw2mvWmxbM/vvD8cdDnz7QvHnAwiIiBVuUqzsXAcdkuz8aWLLTa6oAfZxNqCkJXOacS/feD4ywXSJJZc4cOCVzNuewYVC7dqDCq1bBli1QujR8+GGgoiIiskOUIe1HoLxzriywGGgGXJv9Bd77sjuunXPvA4MV0ESypKfbSQIA778fMKCtWGHHPBUqBD/9BIULByosIiI7RBbSvPfpzrn/YKs2CwM9vPcznHPtM3+f4zw0kYIuI8Omgo0aZXkp2FSwpUuhRg344w/44gsFNBGRfBLpZrbe+yHAkJ1+tstw5r1vHWVbRJJJRgZUrAjTp8OJJ9q2G0EsWgSXXgpLlsBXX0H16oEKi4jIznR2p0iCWboUypWzgHb44TBrVsAjMW+7DZYts8lvCmgiIvlKx0KJJJBZs2ykcUnmEpslSwKfWd69OyxcCGedFbCoiIjsinrSRBLE5Mlw8skWzB5/HLwPNB1szhxo3x62bbOuOwU0EZGEoJAmkgBGjszKRj17whNPBCr86682rDlgAPz5Z6CiIiISCw13iuSzjIyszWkHDIArrwxUeNo0G1stXNiWkJ5wQqDCIiISC/WkieSj9HQ7BnP7dpuzHyygTZkCl1wCRYvC6NFZu+WKiEjCUEgTyUevvAIffAAlS8LrrwcsnJ5uJwmMHm17fIiISMLRcKdIPpk/Hx54wK6XLrXN/SO3eLGFs6pVrTctSFEREdkT+gstkg+8hwsvtOsePaBIiH8ujRkDJ50E77xj9wpoIiIJTX+lRfJBy5bWqdWkCdxwQ4CCI0faGVPHHAMNGgQoKCIie0shTSSwZ56Bjz+GffeFXr0CFBw61IJZuXK2ivOoowIUFRGRvaU5aSKBZGTAv/9tw5sA69YFGOZcvNiWjJ58Mnz9NRx2WMQFRUQkXhTSRALYsgWqVIEZM+x+wQLb/SJypUvDRx/ZfmglSgQoKCIi8aLhTpGIrVgBFSpYQKtQAbZuhTJlIi7at6/NQwNo3FgBTUQkCSmkiUQoLQ3Kl7cTl44/HmbOhH32ibjohx/CtdfCyy/bMlIREUlKCmkiEcnIgFq1bO5Z8+bw++/gXMRF330XWreGiy+GTz8NUFBERKKikCYSkTvvhLFj4dhjA63i7NoVbrzRDgIdPBiKFQtQVEREoqKQJhKBAQOgUye7nj8/QEHvYdIkaNgQBg6E/fcPUFRERKKk1Z0icTZoEFx9tV3Puy2C9gAAIABJREFUmBFgxHHdOjjoIDtJYPv2QMtGRUQkaupJE4mj996Dyy+360mT4JRTIi749NNQqZItIS1cWAFNRCSFKKSJxIH3UL8+tGlj999+C2edFXHBRx+Fxx6DCy7QJrUiIilIw50ie2n7djj3XPjxR7ufPx+OOy7Cgt5Dhw7w0kvQti107269aCIiklLUkyayFzZtgtq1LaCVKGEb1UYa0ADeeMMC2s03w9tvK6CJiKQo9aSJ7KF166ByZZg3z458mjABCoX4Z8/111v33d13ax80EZEUpp40kT2waROceqoFtEaNrCct0oC2fTu8+aYdAnroof/X3p3H+1Ttfxx/fYwnQyrNJCSKwpXGW1QUGpAypKvSoDSaKr+uE0m3QjMNaFZUQoi4mRJ1TSFNh4t04qakMh0dzvr9sb446XC+55zv/o7v5+Phcfb+fvdea3X2Q71ba6+1oGdPBTQRkSSnkCZSQJ9/7teJzcyEDh3g/fcDrnDnTujcGe6+G957L+DKREQkXmi4U6QA1q2DBg388dVXR2EngexsP7w5erRfbuOaawKuUERE4oVCmkiYVq3yS5IBvPmm38M8UH/84ZPg2LEwcCDcc0/AFYqISDxRSBMJQ2YmnHCCPx40KAoBDWDNGpg1C556yg91iohISlFIE8lHZiZUr+6Pb7gBevUKuMLsbChRAmrWhIwMLVQrIpKiNHFAJB9du/rc9NBD8NJLAVe2dSs0bw4DBvhzBTQRkZSlkCZyAC1bwqRJ0KQJ9OkTcGWbN0OLFn6Is1q1gCsTEZF4p+FOkTw45zu0pk3z5x98EHCFv/3mA9r8+X7KaPv2AVcoIiLxTiFNJA+vvuoDWunSvoOrZMkAK9u1yyfCRYvgnXegTZsAKxMRkUShkCayj9de8xMEADZuDDiggd97s2tXv/nn5ZcHXJmIiCQKhTSREOf85IAHH/Tnb73ldxYIzIYNsHw5XHihX7BWREQkF4U0EXxA69bNb48JMHmyf0UsMOvX+9kIP/4Iq1fDwQcHWJmIiCQihTRJec7Bfff5gGbmV8E46KAAK8zM9L1n69f7GQkKaCIikgeFNElp2dlw/vkwbx4ccojfmzPQgLZmjQ9oGzfC1KlwzjkBViYiIolM66RJysrO9iOO8+bBuefC//4XcEADePll2LQJPvpIAU1ERA5IIU1SVvv2MGeOD2hz5vjlNgLjnP/Zrx8sXgynnx5gZSIikgwU0iQlffghjBsHVav6gBaor76CM8/0EwSKFdNuAiIiEha9kyYpZ8MGaNXKHy9aFHBly5ZB06Z+LbSsrIArExGRZKKeNEkpO3ZA69bwxx8wYgQcdliAlS1eDBdcAKVKwezZcPLJAVYmIiLJRj1pkjLWrYNKlfzx5ZfDjTcGWNmSJX4W5yGHwIwZUL16gJWJiEgyUk+apIRt2+Bvf/PHNWrA++8HXGG1an4/ztmzFdBERKRQFNIkJTzwgH8XrU0bWLHCL1obiIULfSKsUAFGj4bjjw+oIhERSXYKaZL0tm6F55+Ho46C994LsKKPPoJGjaBXrwArERGRVKF30iSpOedncm7bBq+8EmBFkyf7brqaNf1aaCIiIkWknjRJWnPn+omV06dDy5bQrl1AFb3/vp8yWqcOzJwJRx4ZUEUiIpJKFNIk6TgHbdv6nQR27oQbboCxYwOqbPt2uP12Pyth+nSoWDGgikREJNVouFOSznPPwZgx/njpUqhbN8DKDjrIv4t27LFw8MEBViQiIqlGIU2Sys8/wx13+OOsrAD343ztNcjIgAED4KSTAqpERERSmYY7Jam0bet/DhwYYEAbPhw6d4b58yE7O6BKREQk1SmkSdJ45RWYNcuvgXbPPQFVMnQodOniF6qdONHPTBAREQmAQpokhZtv9hMEKlSALVsCquSpp/xYaqtWMG4cpKUFVJGIiIhCmiSBKVP8ZukAX38NZcoEVNExx0D79vDuuwGOpYqIiHgKaZLQMjLgkkugRAn48UefoyLKOfjmG3/cvj2MGgUlS0a4EhERkb9SSJOE9dtvUK+eP3711QDWkHUO0tP9Gh6LF/vPAtv0U0RE5M+0BIckpJwcuOkmv8zG/ffDNddEuALn4N57YfBg/8Jb/foRrkBEROTAFNIk4TgHhx3me9IaNYKHHw6ggm7d4Jln/G4CzzwDxdTpLCIi0aX/8khC2b7dv4P222/+fOrUACoZO9YHs+7d4dlnFdBERCQm1JMmCcM5uOgiv3H6IYf43QWKFw+gojZt/Kbpl1+ud9BERCRm1EUgCeO223xAa90aNm2KcEDbudMPcWZk+GDWsqUCmoiIxJR60iQhdOoEI0dCnTp+NDKisrPhH/+Ad96BE06AmjUjXIGIiEjBKaRJ3Bszxge0UqVgwYIId3D98Qd06OB3EBg8GO68M4KFi4iIFJ5CmsS1KVN8L1q5cn5N2YMOimDhWVlw1VXwwQd+ooACmoiIxBG9kyZxa+RIP5MzJwc++wwqVYpwBbt2we+/wwsvKKCJiEjcUU+axKXHHoPevf3x0qVw0kkRLHzrVp/8ypeHmTMDmiIqIiJSNAppEndGjdob0BYujHBA27wZLr3Ub/Y5fboCmoiIxC2FNIkrX38NHTv641WroFq1CBb+66/QooWfffDmm1piQ0RE4preSZO48e9/Q+3a/njChAgHtF9+8SvhLloE774L7dtHsHAREZHIU0+axIWsLGjWzB8PHeoX+4+oTp1g2TK/yNpll0W4cBERkchTSJOYc853bDnn15S97bYAKnn8cfj+e9+bJiIikgA03Ckx5Rx06eKHN088Ed54I4KFr1vnp4k652cfKKCJiEgCUU+axMyOHXD++X4NtGLFYPHiCBb+/fdw4YXwv//5BWtPOCGChYuIiARPPWkSEzt3wl13+YBWvrwPbOXKRajw1auhUSPYsAGmTVNAExGRhKSQJjExaBAMGwbnnecX/S8RqT7dlSuhcWP47Te/DtrZZ0eoYBERkejScKdE3SefwIMPQsWKMGNGhAv/9lvfTTdjBtSvH+HCRUREokchTaJq2TK4+GI/vLlgQQR70LZuhbJl/W4CK1dCmTIRKlhERCQ2NNwpUbNpkx993L4dJk+GU0+NUMFLl0KNGjB+vD9XQBMRkSSgkCZR869/wbZt8OSTfnemiFi4EC64wHfJ1akToUJFRERiTyFNomLUKBg8GI4+Grp1i1Chn30GTZpAhQrw8cd+oTUREZEkoZAmgfvkE79puhnMnRuhQtes8YvTHnEEzJ4d4Y0+RUREYk8hTQL1xRd+mQ3wuwpUrx6hgo8/Hvr29QGtSpUIFSoiIhI/FNIkMIsWwWmn+eP+/SO0r/lHH8GXX/puuV69oFKlCBQqIiISfxTSJBBjxkDDhpCdDUOHQnp6BAr94AO/xEbPnhEoTEREJL5pnTSJuLFjoW1bf/zOO3uPi2T8eGjXDurWhbfeikCBIiIi8U0hTSLqiSf2dnR9+SXUrh2BQt991888OO00+PBDOOSQCBQqIiIS3zTcKRGzYMHegPbFFxEKaM7BiBFw1ll+s3QFNBERSRHqSZMicw4eewz+7//8+aRJcMopESh4506/SO3Ysb6ScuUiUKiIiEhiUEiTInEOjjwSfv7Zn3/4ITRrFoGChw2DV1+FqVOhfPkIFCgiIpJYNNwpRZKe7gPakUfC779HKKANGQK33AKHHQYlS0agQBERkcSjkCaFtnQpPPywP167NkIdXk88AXfeCa1b+2HOtLQIFCoiIpJ4FNKkUJyDu+/2x8uWQenSESj0uef8zIO2bf3aHaVKRaBQERGRxKSQJoXy7LN+R6ZzzoFTT41Qoc2a+d3X33pLw5wiIpLyFNKkwHJy9u4gMGtWEQtzzi9Um5MDJ5wATz7pZ3SKiIikOIU0KZBdu+DCC/0kgVtuKWKHl3N+/80rrvAL1oqIiMge6rKQAunTxw9zVqniJ2EWWk6Of6ltyBA/UaBdu4i1UUREJBkopEnYhg+HRx/1x999V4SCcnKga1e/FlrPnjBoEJhFpI0iIiLJItDhTjNrbmbfmtlKM+udx/fXmNmy0J95ZlYvyPZI4WRn+zzVpYs/nz+/iAV+8YVfqPb++xXQRERE9iOwnjQzKw4MBS4CMoEFZjbBOfdVrstWA42dc5vMrAUwDDgzqDZJ4dx4I7zxhj+eOBFOP72QBTnnA1m9en7djpo1FdBERET2I8ietDOAlc65Vc65P4DRQKvcFzjn5jnnNoVOPwMqB9geKaAdO6BVq70BLSsLLruskIVlZ8PVV8PIkf68Vi0FNBERkQMIMqRVAr7PdZ4Z+mx/bgSm5PWFmXUxs4VmtjCC7ZMDcA46doQJE/z5tm1FWLB2xw6/QO3bb8OGDRFro4iISDILMqTl1U3i8rzQ7AJ8SLsvr++dc8Occw2dcw0j2D45gPR0vytTq1b+Pf+DDipkQVlZ0KYNvP++XwG3R4+ItlNERCRZBTm7MxM4Ltd5ZWDdvheZWV1gBNDCObcxwPZIGJyDfv327sk5dmwRRiWzs6FlS/joI3jxxb0zD0RERCRfQYa0BcCJZlYN+AHoAHTMfYGZVQHGAp2ccxkBtkXC1LWrz1MAP/4IxYrS11qypN83qmNHuP76SDRPREQkZQQW0pxzO83sDmAqUBx42Tn3pZndGvr+BeABoCLwnPnump0a0oydTp38e/0HHwwrV8IRRxSyoN9/h++/hzp1fLeciIiIFFigi9k65yYDk/f57IVcxzcBNwXZBglPmzYwbhw0aADz5hVhksCvv0Lz5n6125UroWzZiLZTREQkVWjHgRTnHFx3nQ9oANOnFyGgbdwIF1/sF6t9910FNBERkSJQSEthK1ZAs2awejUcdpjv/CpXrpCFbdgAF10E334L48fDJZdEtK0iIiKpJtBtoSR+/fCDX/B/9Wpo3dq/QlbogAYwYIBPfRMnKqCJiIhEgEJaitq9UfrNN/uhzjJliljgwIHw8ce+N01ERESKTCEtBb35JgwZApUr+43TC23tWr+TwKZNkJYGDTUxV0REJFL0TlqK+eADuPZafzxzZhEKWr0aLrjAz+ZcswYOPTQSzRMREZEQ9aSlkPXr4Zpr/DZP8+ZBjRqFLGjFCmjUCDZv9tNB//a3iLZTRERE1JOWMlas8BMFwG+jefbZhSzom2/gwgv9lk8zZkC9ehFro4iIiOylnrQUkJUF9ev74169/HaahVamDFStCrNmKaCJiIgESD1pKeCee2DbNujWDQYNKmQhq1bB8cdDlSowd24Rdl0XERGRcKgnLcmtWeNnclavDk8+WchCFi70Mzf79PHnCmgiIiKBU0hLYmPH7n0PbcyYQhby6afQpAlUqAC33BKxtomIiMiBKaQlqbffhiuv9O/3P/10ISdgzpnj9+I88ki/UG3VqpFupoiIiOyH3klLQj/9BLfd5jdK37ABDj64EIVs2QJt2vgVb6dPh2OPjXg7RUREZP8U0pLQ9dfDL7/4d9EKFdDAb+Q5ZgycdBIcdVQkmyciIiJh0HBnknnkEZg8GTp0gNtvL0QBkybBSy/548aNFdBERERiRCEtiUyZAvffD6VKweOPF6KAceP8EOewYbBzZ8TbJyIiIuFTSEsSDzwAl1zij//730K8Qvb2236z9IYNYdo0KKGRcBERkVhSSEsCH3wADz3kj2fN8u/6F8jIkdCxI5xzDkyd6pfbEBERkZhSd0mCmzzZL7UBsHQp1K1biEIyM+H882HCBChbNpLNExERkUJST1oCe+kluPRS2LEDFi0qRED7+Wf/s3dv34OmgCYiIhI3FNIS1MSJcNNN/njOHGjQoIAFPPMMnHgifPONP9c7aCIiInFF/2VOQGvWwK23+uO1a+G44wpYwKBBcO+9fiZn9eqRbp6IiIhEgHrSEsy6ddCokf85bFghAtrDD/uA1r49jB7t1+sQERGRuKOetASSkeE3AHAO0tPh5psLWMDo0dCnD3TqBC+/rCFOERGROKaetATx4otQq5YPaPfcA/37F6KQNm3g2WfhlVcU0EREROKcQloCyMjY+w7a88/DwIEFuNk5GDzYz+QsVQruuAOKFw+knSIiIhI56k6Jc5s3wwUX+ONPP4WzzirAzTk5cNddMHQomEHPnoG0UURERCJPIS2OOQfNm/tJAv36FSKg3XILjBjhx0d79AiqmSIiIhIADXfGse7dYd48v0vTAw8U4MZdu+CGG3xA69MHHnvM96SJiIhIwlBPWpyaMgWeftofr1tXwIy1aZNPd/37+2mgIiIiknAU0uLQL7/Abbf5440boUyZMG/MzvZp7vDD/T5R5csH1kYREREJloY741B6ut9VYNAgOOywMG/asQOuugo6d/YvsymgiYiIJDSFtDjz+efw3HNwzjnQq1eYN23fDldcARMmwNln6/0zERGRJKDhzjize9P0J54I84Zt26BVK5g+HYYP31uAiIiIJDSFtDiRnQ2nnQZffAHXXgtnnhnmje3bw4wZ8Oqr/kYRERFJCgppceKqq3xAA7+rQNh69YJrroEOHQJpl4iIiMSG3kmLAytWwMSJfm/OnJwwZnNu2gRvv+2PGzdWQBMREUlCCmkxtnYtNGrkJ2SOGxfGO/8bN0KTJnDddf5mERERSUoa7oyh33+H00+HDRtg7Fg4+eR8btiwAZo29TuujxsHVapEpZ0iIiISfQppMdS/v89dr73mV9A4oPXrfQ/amjUwaZIPayIiIpK0FNJi5N134fHH4YQTwpyUOXWqH96cMsW/hyYiIiJJTSEtBmbMgHbt/PGbb+ZzcU4OFCsG118PzZrBMccE3TwRERGJA5o4EGXdu/tRS4Avv8xnPbRVq6BePfj0U3+ugCYiIpIy1JMWRY88Ak895Y+nTYPatQ9w8YoVcMEFfsun0qWj0j4RERGJHwppUfLss3D//X6JjR07oGTJA1z89ddw4YWwaxfMnAl160atnSIiIhIfNNwZBQsWwF13QcWKkJmZT0BbvXrvxIBZsxTQREREUpRCWsBycvzOTeAnDBx7bD43VK4MbdvC7Nn5jIeKiIhIMtNwZ4B27fI5KyMDOnXKp1Ns0SKf4I45BoYOjVobRUREJD6pJy1At97qA1qLFn7B2v2aN89PEujSJWptExERkfimkBaQiRNhxAi/7dPkyQfYk3P2bLj4Yjj6aHj++ai2UUREROKXQloAnPMzOQFeeOEAF06f7rvZqlTxYa1y5ai0T0REROKfQloA7r8fli+He+6BBg32c1FODtx3H9So4WdxaqFaERERyUUTByLsmWfg0UeheHF48MEDXFismN8ovUQJOPzwqLVPREREEoN60iJo+HDo1g0OPRR++w0OOiiPi957Dzp2hJ07/XtoCmgiIiKSB4W0CBkyxE/OdA7mzoWyZfO4aPRoaN8e1qzx2z2JiIiI7IeGOyNg8WK4805IS/M7ClSsmMdFr78OnTvDuef6Yc7y5aPeThEREUkc6kkrotmzoWlTfzxr1n4C2muvwfXXw/nn+/U4FNBEREQkHwppRfDddz53bdoEH3wAZ565nwtr1YKrrvI9aHmOg4qIiIj8mUJaIf3yy94etE8/hUsuyeOixYv9z7POgnfe2c9MAhEREZG/UkgrpNNPh5Ur4fbbfQb7i4ED4bTTfO+ZiIiISAEppBXCU0/BqlV+u80hQ/K44KGH/EK1HTpA8+ZRb5+IiIgkPoW0AurZE7p398fTpu3zpXOQng4PPACdOsHIkX6xWhEREZECUkgrgHnz4Ikn/G4CS5bkkb/mz4cBA+DGG+GVV/yFIiIiIoWgbp4wzZgBrVr5488+g3r18rjozDNh5kxo1Mhv+yQiIiJSSEoSYfjuOz+Tc8sWmDIFGjbM9WVODvTo4RdMA78mhwKaiIiIFJF60vLhHLRt63/OmOEnC+yxaxfccgu89BKUKweNG8esnSIiIpJcFNLycf/9sGABXHfdPgFt50644QZ44w0/WeDBB2PWRhEREUk+CmkHsHAhPPooHH207yzbIzvbz958+22/3EafPjFro4iIiCQnhbT9yMjw8wBKlYK5c/eZqFmsGJQs6ResveeemLVRREREkpdCWh5+/RUuvNDPCXjvPahePfTFjh1+P6hjjoHXXwezmLZTREREkpemIe5j3Tq/vMYPP0CXLnDZZaEvtm/3a3Ccfz5kZSmgiYiISKDUk5bLzz9D3bqwcaNfk/af/wx9sXUrtGzp10AbPhzS0mLaThEREUl+Cmm5pKf7gJaeniugbd4Ml17qX0x77TU/YUBEREQkYAppIenp8MILcOqp0L9/ri969PD7Qb31FrRvH7P2iYiISGpRSMNPDhgwwB9PnLjPl488AldeCc2bR71dIiIikrpSfuLAxx9Du3b+eMMGOP54/MtpvXrBH3/A4YcroImIiEjUpXRI+/VXv5OTczBrFhxxBPDjj35rgaFDYenSWDdRREREUlTKDndu2QJ//7s/fuGF0Lab69f7BdK++w4mTYLTT49pG0VERCR1pWRP2kcfQfny8NVXcN55fj00MjN9UsvMhA8/hCZNYt1MERERSWEp15O2bBlcdJE/vv12GDIk9MWGDX5HgWnT4OyzY9Y+ERGRosrOziYzM5OsrKxYNyVlpKWlUblyZUqWLBmxMlMqpO3c6TcMAJgwAS6/HNi0CQ49FBo0gBUr/GadIiIiCSwzM5Py5ctTtWpVTDvkBM45x8aNG8nMzKRatWoRKzelhjvbtvWZ7OqrQwHt22/hlFPgqaf8BQpoIiKSBLKysqhYsaICWpSYGRUrVox4z2VKhLT//te/ezZ+PJQoAW++iX8hrXFj373WtGmsmygiIhJRCmjRFcTvO+lD2s03Q40a8MknULq0n8BpXyzz457Fivm1N045JdbNFBEREfmTpA5pffrAiBH+ePRoyMqCw0v+5mduli4Ns2fDySfHtpEiIiJJaty4cZgZ33zzzZ7PZs2axWWXXfan666//nrGjBkD+EkPvXv35sQTT+SUU07hjDPOYMqUKUVuyyOPPEKNGjWoVasWU6dOzfOafv36UalSJerXr0/9+vWZPHnyn75fu3Yt5cqVY/DgwUVuTziSduLA2LEweDAccojfQKB48dAXFSrA00/DWWdB9eoxbaOIiEgyGzVqFOeeey6jR4+mX79+Yd2Tnp7O+vXrWb58OaVLl+bHH39k9uzZRWrHV199xejRo/nyyy9Zt24dTZs2JSMjg+J7wsFe3bt3p1evXnmW0717d1q0aFGkthREUoa02bP9dpsAGRmhgDZ3LmzdChdfDB07xrR9IiIi0dKtGyxZEtky69ffO+duf7Zs2cLcuXOZOXMmLVu2DCukbdu2jeHDh7N69WpKly4NwFFHHUW73fs3FtL7779Phw4dKF26NNWqVaNGjRrMnz+fswuw5Nb48eOpXr06ZcuWLVJbCiLphjudg7vv9scrV0KVKvj3zpo1g/vug5ycWDZPREQkJYwfP57mzZtTs2ZNDjvsMBYvXpzvPStXrqRKlSocfPDB+V7bvXv3PcOSuf88+uijf7n2hx9+4LjjjttzXrlyZX744Yc8yx0yZAh169blhhtuYNOmTQBs3bqVxx57jL59++bbrkhKup60rl39lpsdO8IJJ+C3F2jZEqpVgylT/GQBERGRFJFfj1dQRo0aRbdu3QDo0KEDo0aNokGDBvudBVnQ2ZFPPvlk2Nc658Kqr2vXrqSnp2NmpKen07NnT15++WX69u1L9+7dKVeuXIHaWFRJFdJ+/BFefx2OPBLeeAMfyq64AmrW9GHtyCNj3UQREZGkt3HjRmbMmMHy5csxM3bt2oWZMXDgQCpWrLinh2q3X375hcMPP5waNWqwdu1aNm/eTPny5Q9YR/fu3Zk5c+ZfPu/QoQO9e/f+02eVK1fm+++/33OemZnJscce+5d7jzrqqD3HN998854JDv/5z38YM2YM9957L7/++ivFihUjLS2NO+64I/9fRhEkVUj75z9h+3aYPj3UYTZxItSp47d6qlgx1s0TERFJCWPGjOHaa6/lxRdf3PNZ48aN+eSTTzjjjDNYt24dX3/9NSeffDLfffcdS5cupX79+pQpU4Ybb7yRu+66ixdffJFSpUqxfv16pk+fzj/+8Y8/1VGQnrSWLVvSsWNHevTowbp161ixYgVnnHHGX65bv349xxxzDOBnpp4SWqJrzpw5e67p168f5cqVCzygQRKFtFWr4KWX/BDn2Q12AKX9xpxbt/rd1EVERCQqRo0a9ZferCuvvJK33nqL8847j5EjR9K5c2eysrIoWbIkI0aMoEKFCgAMGDCAPn36ULt2bdLS0ihbtiz9+/cvUnvq1KlDu3btqF27NiVKlGDo0KF7ZnbedNNN3HrrrTRs2JB7772XJUuWYGZUrVr1TyEzFiyvcdp4Vr91a7dk/Pg/feYcnHsuzJsH3z/2FpWH94WZM6Fy5Ri1UkREJHZ291JJdOX1e6/fujVLxo8v1HYESfEW/aRJPqANOeM1Kvf+B1Sq5BdIExEREUlQCR/SMjKgdWu4tfhwblvQ2e8mMHkyRHkGhoiIiEgkJXRI27kTatWCK3Pe4fldXbDmzf1kgTJlYt00ERGRmEq015kSXRC/74QNaTk50LSpP651+0Vw770wbhykpcW2YSIiIjGWlpbGxo0bFdSixDnHxo0bSYtwBknY2Z39+sHRs0dz+UWt6fvUoVDisVg3SUREJC5UrlyZzMxMfvrpp1g3JWWkpaVROcITFhMypA0bBjzUn9H0xV08GCvRM9ZNEhERiRslS5akWrVqsW6GFFGgw51m1tzMvjWzlWbWO4/vzcyeCX2/zMwahFPuth596E9f/uh4Hda9W+QbLiIiIhJjgYU0MysODAVaALWBq82s9j6XtQBODP3pAjwfTtndtj7MqiY3U+qNlyG0GJ2IiIhIMgmyJ+0MYKVzbpUJ+UQkAAAIFklEQVRz7g9gNNBqn2taAa877zPgEDM7Jr+C3zz0dqpPe0GbpYuIiEjSCvKdtErA97nOM4Ezw7imErA+90Vm1gXf00a9Vq0Y1CiTQW3aRLzBIiIiIvEiyJCW1xYI+84FDucanHPDgGEAZrbQOdew6M2TWNDzS1x6dolNzy+x6fklLjNbWNh7gxwvzASOy3VeGVhXiGtEREREUk6QIW0BcKKZVTOzUkAHYMI+10wArg3N8jwL+M05t37fgkRERERSTWDDnc65nWZ2BzAVKA687Jz70sxuDX3/AjAZuARYCWwDOodR9LCAmizRoeeXuPTsEpueX2LT80tchX52pi0jREREROKP1rAQERERiUMKaSIiIiJxKG5DWlBbSknwwnh214Se2TIzm2dm9WLRTslbfs8v13Wnm9kuM7sqmu2TAwvn+ZnZ+Wa2xMy+NLPZ0W6j5C2Mf3dWMLOJZrY09OzCeY9bosDMXjazDWa2fD/fFyqzxGVIC3JLKQlWmM9uNdDYOVcXeAi9EBs3wnx+u697DD8xSOJEOM/PzA4BngNaOufqAG2j3lD5izD/7t0OfOWcqwecDzweWj1BYu9VoPkBvi9UZonLkEaAW0pJ4PJ9ds65ec65TaHTz/Dr40l8COfvHsCdwHvAhmg2TvIVzvPrCIx1zq0FcM7pGcaHcJ6dA8qbmQHlgF+AndFtpuTFOfcx/nnsT6EyS7yGtP1tF1XQayT6CvpcbgSmBNoiKYh8n5+ZVQKuAF6IYrskPOH8/asJHGpms8xskZldG7XWyYGE8+yGACfjF33/ArjbOZcTneZJERUqswS5LVRRRGxLKYm6sJ+LmV2AD2nnBtoiKYhwnt9TwH3OuV3+f+gljoTz/EoApwFNgIOAT83sM+dcRtCNkwMK59k1A5YAFwInAP82sznOud+DbpwUWaEyS7yGNG0plbjCei5mVhcYAbRwzm2MUtskf+E8v4bA6FBAOxy4xMx2OufGR6eJcgDh/rvzZ+fcVmCrmX0M1AMU0mIrnGfXGXjU+QVOV5rZauAkYH50mihFUKjMEq/DndpSKnHl++zMrAowFuik/3uPO/k+P+dcNedcVedcVWAMcJsCWtwI59+d7wPnmVkJMysDnAl8HeV2yl+F8+zW4ntAMbOjgFrAqqi2UgqrUJklLnvSAtxSSgIW5rN7AKgIPBfqjdnpnGsYqzbLXmE+P4lT4Tw/59zXZvYhsAzIAUY45/JcNkCiJ8y/ew8Br5rZF/jhs/uccz/HrNGyh5mNws+4PdzMMoG+QEkoWmbRtlAiIiIicShehztFREREUppCmoiIiEgcUkgTERERiUMKaSIiIiJxSCFNREREJA4ppIlIxJnZLjNbkutP1QNcuyUC9b1qZqtDdS02s7MLUcaI3Rtam9n9+3w3r6htDJWz+/ey3MwmhjY7P9D19c3skkjULSKJR0twiEjEmdkW51y5SF97gDJeBSY558aY2cXAYOdc3SKUV+Q25Veumb0GZDjnHj7A9dcDDZ1zd0S6LSIS/9STJiKBM7NyZjY91Mv1hZm1yuOaY8zs41w9TeeFPr/YzD4N3fuumeUXnj4GaoTu7REqa7mZdQt9VtbMPjCzpaHP24c+n2VmDc3sUeCgUDveDH23JfTz7dw9W6EevCvNrLiZDTKzBWa2zMxuCePX8imhDZbN7Awzm2dmn4d+1gqtOt8faB9qS/tQ218O1fN5Xr9HEUkecbnjgIgkvIPMbEnoeDXQFrjCOfe7mR0OfGZmE9yfu/I7AlOdcw+bWXGgTOjaPkBT59xWM7sP6IEPL/tzOfCFmZ2GX9X7TPzq7P8xs9lAdWCdc+5SADOrkPtm51xvM7vDOVc/j7JHA+2ByaEQ1QToCtyI3+bldDMrDcw1s2nOudV5NTD0z9cEeCn00TdAo9Cq802BfznnrjSzB8jVk2Zm/wJmOOduCA2Vzjezj0L7cIpIklFIE5EgbM8dcsysJPAvM2uE34qoEnAU8L9c9ywAXg5dO945t8TMGgO18aEHoBS+Byovg8ysD/ATPjQ1AcbtDjBmNhY4D/gQGGxmj+GHSOcU4J9rCvBMKIg1Bz52zm0PDbHWNbOrQtdVAE7EB9TcdofXqsAi4N+5rn/NzE4EHKHtZPJwMdDSzHqFztOAKmjvTZGkpJAmItFwDXAEcJpzLtvM1uADxh7OuY9DIe5S4A0zGwRsAv7tnLs6jDrucc6N2X0S6pH6C+dcRqiX7RLgkVCP14F65nLfm2Vms4Bm+B61UburA+50zk3Np4jtzrn6od67ScDtwDP4PRlnOueuCE2ymLWf+w240jn3bTjtFZHEpnfSRCQaKgAbQgHtAuD4fS8ws+ND1wzHDwM2AD4D/m5mu98xK2NmNcOs82OgdeiessAVwBwzOxbY5pwbCQwO1bOv7FCPXl5G44dRz8Nvhk3oZ9fd95hZzVCdeXLO/QbcBfQK3VMB+CH09fW5Lt0MlM91PhW400Ldimb2t/3VISKJTyFNRKLhTaChmS3E96p9k8c15wNLzOxz4ErgaefcT/jQMsrMluFD20nhVOicWwy8CswH/gOMcM59DpyKf5drCfBPYEAetw8Dlu2eOLCPaUAj4CPn3B+hz0YAXwGLzWw58CL5jFSE2rIU6AAMxPfqzQWK57psJlB798QBfI9byVDblofORSRJaQkOERERkTiknjQRERGROKSQJiIiIhKHFNJERERE4pBCmoiIiEgcUkgTERERiUMKaSIiIiJxSCFNREREJA79P9syD2czfquTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['6:32651641' '6:32438298' '6:29933439' '11:10490415' '6:26364628'\n",
      " '5:96124330' '1:150861859' '11:10489826' '6:32634243' '5:96216173'\n",
      " '5:96115146' '5:95981776' '1:150558293' '6:29910358' '5:96229272'\n",
      " '6:31322980' '6:32204433' '6:32602396' '6:32635629' '6:32212241'\n",
      " '6:32647701' '6:32575658' '6:32659609' '6:32134510']\n"
     ]
    }
   ],
   "source": [
    "# define model\n",
    "model = Lasso(alpha=0.00001,random_state=random_sate)\n",
    "# fit model\n",
    "model.fit(SNP_outcome_cov_table_training_x, SNP_outcome_cov_table_training_y)\n",
    "p = model.predict(SNP_outcome_cov_table_training_x)  \n",
    "preds=p\n",
    "#Find model that gives best AUC, do a cutoff of weakest contributing SNPS, then run all analysis port../tables/reds=p\n",
    "fpr, tpr, threshold = metrics.roc_curve(list(SNP_outcome_cov_table_training_y), preds,pos_label=1)\n",
    "roc_auc = metrics.auc(fpr, tpr)\n",
    "print(roc_auc)\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.title('Simple LASSO AUC')\n",
    "plt.plot(fpr, tpr, 'b', label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],'r--')\n",
    "plt.xlim([0, 1])\n",
    "plt.ylim([0, 1])\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()\n",
    "\n",
    "#now drop SNPs with low contribution\n",
    "print(SNP_outcome_cov_table_training_x.columns.values[model.coef_==0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "5983c027-c1a9-4d08-ae36-d04b5f6260b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "#remake SNP list with dropped \"useless\" SNPs\n",
    "\n",
    "ellipse_outcome_cov_table=ellipse_outcome_cov_table.drop(SNP_outcome_cov_table_training_x.columns.values[model.coef_==0],axis=1)\n",
    "SNP_outcome_cov_table_downsamp=SNP_outcome_cov_table_downsamp.drop(SNP_outcome_cov_table_training_x.columns.values[model.coef_==0],axis=1)\n",
    "\n",
    "ellipse_outcome_cov_table_x=ellipse_outcome_cov_table.drop(['IID',\n",
    "       'PRS', 'pheno', 'age',\n",
    "       'FH_PROS', 'FID', 'PC1', 'PC2', 'PC3', 'PC4',\n",
    "       'PC5', 'PC6', 'PC7', 'PC8', 'PC9', 'PC10'],axis=1)\n",
    "ellipse_outcome_cov_table_x.index=ellipse_outcome_cov_table[\"IID\"]\n",
    "ellipse_outcome_cov_table_x=ellipse_outcome_cov_table_x.drop([\"index_x\",\"index_y\"],axis=1)\n",
    "\n",
    "ellipse_outcome_cov_table_y=ellipse_outcome_cov_table[\"pheno\"]\n",
    "ellipse_outcome_cov_table_y.index=ellipse_outcome_cov_table[\"IID\"]\n",
    "\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_downsamp.drop([\"IID\",\"Cancer\",\"C43-C44\",\"C60-C63\",\"AGE\",\"SEX\",\"PC1\",\"PC2\",\"PC3\",\"PC4\",\"PC5\",\"PC6\",\"PC7\",\"PC8\",\"PC9\",\"PC10\",'PRS',\n",
    "       'FID', 'D55-D59', 'D80-D89', 'E70-E90', 'I70-I79', 'K20-K31',\n",
    "       'K50-K52', 'K70-K77', 'K80-K87', 'K90-K93', 'L40-L45', 'M00-M03',\n",
    "       'M05-M14', 'M15-M19', 'M20-M25', 'M70-M79', 'M45-M49', 'M60-M63',\n",
    "       'R70-R79', 'C73-C75', 'C81-C96', 'D60-D64', 'C40-C41', 'D37-D48',\n",
    "       'D70-D77', 'C50-C50', 'D65-D69', 'C00-C14', 'C45-C49', 'D00-D09',\n",
    "       'C30-C39', 'C97-C97', 'D50-D53', 'C76-C80', 'C15-C26', 'C51-C58',\n",
    "       'D10-D36', 'C69-C72', 'C64-C68'],axis=1)\n",
    "SNP_outcome_cov_table_training_x.index=SNP_outcome_cov_table_downsamp[\"IID\"]\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_training_x.drop([\"level_0\",\"index\"],axis=1)\n",
    "SNP_outcome_cov_table_training_y=SNP_outcome_cov_table_downsamp[\"Cancer\"]\n",
    "SNP_outcome_cov_table_training_y.index=SNP_outcome_cov_table_downsamp[\"IID\"]\n",
    "\n",
    "ellipse_outcome_cov_table=ellipse_outcome_cov_table.reset_index()\n",
    "\n",
    "SNP_outcome_cov_table_training_x=SNP_outcome_cov_table_training_x[ellipse_outcome_cov_table_x.columns.values]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "28a20460-bcc8-45d8-b667-59d3f9b0a4a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.DataFrame(ellipse_outcome_cov_table_x.columns.values).to_csv(\"../data/risk_model/prostate_final_snps.txt\",sep=\"\\t\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "b7bbe455-07cd-4bf7-a38a-87850962ffcc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[09:27:45] WARNING: /tmp/abs_40obctay9q/croots/recipe/xgboost-split_1659548945886/work/src/learner.cc:576: \n",
      "Parameters: { \"verbose_eval\" } might not be used.\n",
      "\n",
      "  This could be a false alarm, with some parameters getting used by language bindings but\n",
      "  then being mistakenly passed down to XGBoost core, or some parameter actually being used\n",
      "  but getting flagged wrongly here. Please open an issue if you find any such cases.\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "regressor_xgb = xgboost.XGBRFRegressor(n_estimators = 200, random_state = random_sate,learning_rate=0.1,max_depth=9,verbose_eval=10)\n",
    "regressor_xgb.fit(SNP_outcome_cov_table_training_x, SNP_outcome_cov_table_training_y)  \n",
    "\n",
    "p = regressor_xgb.predict(ellipse_outcome_cov_table_x)  # test the output by changing values\n",
    "p_disc=regressor_xgb.predict(SNP_outcome_cov_table_training_x)\n",
    "\n",
    "ellipse_outcome_cov_table[\"LASSO_burden\"]=p\n",
    "SNP_outcome_cov_table_downsamp[\"LASSO_burden\"]=p_disc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "70c80610-f8bd-44bd-8e57-3d9b4b900d9a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAikAAAHbCAYAAAAOIQqRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzcf5xdVX3v/9f7zDDMJJNMfjAhyqAhJaAIJJCIxl6M4q9IFYKAhq+1+INSKWDlFss31fTWpl4FqdobvrRSNIXvbR1UMGCBoOaSaOtEzQmBBMIvY5AxJAQSJswkk5Mz53P/OGvg5DhJJj9mZid5Px+P85h91tpr7bX3Y8h+s/bao4jAzMzMLGtyQz0AMzMzs744pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJtUO9QAOB1NmzYqVCxdqD7v4PW8zMzuS7Ome2G+eSTEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0waspAi6WJJj0oqSZpWVTdH0tOSnpD0voryJalsZfqMS+WflrQqlf2npFNS+Tsr9l0pqVvSrFT3LUkPS3pE0vclNabyj6ayRyT9XNLkwbsqZmZ2qCoUiixtW0fr3atY2raOQqE41EM65NUO9AEkCVBElKqqVgMfAr5Ztf8pwGzgTcBrgZ9IOikietIuH42I5VV9/XtE/HNqfx7wNWBmRDwITEnlY4CngR+lNtdExNZU9zXgKuArwG+AGRGxRdL7gVuAtxzINcjn8+Tz+QPpwszMMmzji0VaF3XQ1V2ivr6BkSMaGTOqgfnzzmXSCWOHeniHrAGZSZE0QdIaSTcDK4DFklan2Y5rACJiTUQ80Ufz84HWiNgREb+hHCzO2tPxesNGMhyIPna7CLg/IrZVtkkhqqG3TUT8PCK2pDbLgJb+nbWZmR2Jij1B66IOduwsMXJ4jnFjGji2uZHObQWunnufZ1QOwEA+7jkZuB24DChGxKkRcRqwYC/tjgOerfjensp6LUiPbuamgAGApCsl/Rq4AfhMH/3OBr5TWSBpAbABeAMwv482nwLu38t4zczsCLa2vUBXd4lh9bveUkc3NbD5pe205duHaGSHvoEMKc9ExDJgLTBR0nxJM4Gte2mnPsp6Z0Y+moLO2enzsVd2iPj/IuIPgOuAL+zSofQa4DTggV06jfgE5UdKa4CPVLV5J+WQcl2fg5Qul7RcUvWjJzMzO4J0vFyiVL2gIenpKbFhU+fgDugwMpAhpQsgPTqZDCwBrgRu3Uu7duD4iu8twPrU1+/Sz5eBf6fvx0CtwKyqsg8DP4iIndU7p7UudwAX9pZJOj2N8/yIeLGvQUbELRExLSKm9VVvZmZHhqYROXK7uZvW1OQY39w4uAM6jAz42z2SjgFyEXEnMBc4cy9N7gFmSzpa0gnAJOCXkmpTX0g6CvgA5cW3SJpU0f6PgKeq+ryEikc9Kjuxdxv4IPB4+v464C7gYxHx5H6cspmZHUEmttQxvD7Htu5dp1O2dGxnzKgGpk/10sb9NeBv91BeT7JAUm8gmgMg6QLK60CagXslrYyI90XEo5K+CzwGFIErI6JH0nDggRRQaoCfAP+S+rxK0ruBncAW4NLeg0uaQHlmZmnFmATcJmlk2n4YuCLV/Q0wFrg5LXkperbEzMx2p7ZGzJ7ZROuiDrZ2ldjZs53unTWvvN1TVzcYt9rDkyL6ehHG9sWUWbNi5cKFfa2l6eWLbGZ2mCsUirTl29mwqZPxzY1Mn9pyJAeUPd0T++2IvXpmZmYHU11dLTOmTxjqYRxW/GfxzczMLJMcUszMzCyTHFLMzMwskxxSzMzMLJMcUszMzCyTHFLMzMwskxxSzMzMLJMcUszMzCyTHFLMzMwskxxSzMzMLJMcUszMzCyTHFLMzMwskxxSzMzMLJMcUszMzCyTHFLMzMwskxxSzMzMLJMcUszMzCyTHFLMzMwsk4YspEi6WNKjkkqSplWUj5X0oKROSTdVlA+TdK+kx1O7r1TUvV7SYkmPSFoiqSWVv1PSyopPt6RZqe5bkh5Obb4vqTGVv0NSR0Wbvxm8q2JmZoeCQqHI0rZ1tN69iqVt6ygUikM9pMNS7UAfQJIARUSpqmo18CHgm1Xl3cBc4NT0qXRjRDwoqQ5YLOn9EXE/cCNwe0TcJukc4MvAxyLiQWBKGscY4GngR6mvayJia6r7GnAV0Bt8fhYRHzigE6+Qz+fJ5/MHqzszMxtCG18s0rqog67uEvX1DYwc0ciYUQ3Mn3cuk04YO9TDO6wMyEyKpAmS1ki6GVhBOVCslrRK0jUAEbEmIp6obhsRXRHxn5TDSmX5thQ6iIhC6rclVZ8CLE7bDwLn9zGsi4D7I2Jb6qM3oAhoAOKATtrMzA57xZ6gdVEHO3aWGDk8x7gxDRzb3EjntgJXz73PMyoH2UA+7jkZuB24DChGxKkRcRqw4EA7ljQK+CCvBpOHgQvT9gXACEnVcXY28J2qfhYAG4A3APMrqqanR0H3S3rTgY7XzMwOD2vbC3R1lxhWv+vtc3RTA5tf2k5bvn2IRnZ4GsiQ8kxELAPWAhMlzZc0E9h6IJ1KqqUcNv5XRKxNxdcCMyQ9BMwAfgcUK9q8BjgNeKCyr4j4BPBaYA3wkVS8Anh9REymHFwW7mYcl0taLmn5gZyPmZkdOjpeLlGqXryQ9PSU2LCpc3AHdJgbyJDSBRARW4DJwBLgSuDWA+z3FuCpiPhGb0FErI+ID0XEGcDnU1lHRZsPAz+IiJ3VnUVED3AHaSYmIrZGRGfavg84StIxfbS7JSKmRcS06jozMzs8NY3IkdvNnbOmJsf45sbBHdBhbsDf7kk3+FxE3El5QeyZB9DX3wNNwGerjyGp91zmAN+uanoJFY96VHZi7zblR0ePp+/jUxmSzqJ8jV7c3zGbmdnhY2JLHcPrc2zr3nU6ZUvHdsaMamD61JbdtLT9MeBv9wDHAQuqQgSSLqD8OKUZuFfSyoh4X6pbB4wE6tIrw++l/Jjo85TDxIqUI26KiFuBdwBflhTATynP2JD6mgAcDyytGJOA2ySNTNsPA1ekuouAKyQVge3A7IjwolozM6O2Rsye2UTrog62dpXY2bOd7p01r7zdU1c3GLfVI4d8/z1wU2bNipULF2oPu/gim5kdRgqFIm35djZs6mR8cyPTp7Y4oOxqT/fEfvMVNTMz20d1dbXMmD5hqIdx2POfxTczM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0wa0pAi6WpJT0h6VNINVXWvk9Qp6dqKskskrZL0iKRFko6panORpJA0LX1/vaS8pJXpGJ+u2PcESb+Q9JSkOyTVpfKPpv4fkfRzSZMH9iqYmR3+CoUiS9vW0Xr3Kpa2raNQKA71kOwQUDvQB5AkQBFRqip/J3A+cHpE7JA0rqrp14H7K/avBf4ROCUiXkih5irgb1P9COAzwC8q+ngOeFvqvxFYLemeiFgPXA98PSJaJf0z8Cngn4DfADMiYouk9wO3AG85kGuQz+fJ5/MH0oWZ2SFr44tFWhd10NVdor6+gZEjGhkzqoH5885l0gljh3p4lmEDMpMiaYKkNZJuBlYAiyWtTrMg16TdrgC+EhE7ACLi+Yr2s4C1wKOV3abP8BR8RgLrK+rnATcA3b0FEVHo7R84mnS+qf05wPdT3W3ArNTm5xGxJZUvA1r2/0qYmR3Zij1B66IOduwsMXJ4jnFjGji2uZHObQWunnufZ1Rsjwbycc/JwO3AZUAxIk6NiNOABan+JODs9MhlqaQ3A0gaDlwHfLGys4jYSTnYrKIcTk4BvpXanAEcHxH/UT0IScdLegR4Frg+zaKMBV6KiN7/OtqB4/o4h09RMZtjZmb7Zm17ga7uEsPqd73djG5qYPNL22nLtw/RyOxQMJAh5ZmIWEZ5RmSipPmSZgJbU30tMBp4K/A54LtphuOLlB/DdFZ2JukoyiHlDOC1wCPAHEk5yo+G/rKvQUTEsxFxOnAicKmkYynPyPzerlXHeyflkHJdX/1KulzScknL93IdzMyOWB0vlyiV+q7r6SmxYVNn35VmDGxI6QJIj04mA0uAK4FbU307cFeU/RIoAcdQXv9xg6R1wGeBv5Z0FTAl9ffriAjgu8DbgBHAqcCS1OatwD29i2d7pRmUR4GzgReAUWmdC5Qf6bzy6EjS6Wmc50fEi32dXETcEhHTImJaX/VmZgZNI3LkdnOnqanJMb65cXAHZIeUAX+7J72Bk4uIO4G5wJmpaiHldSFIOgmoA16IiLMjYkJETAC+AfzPiLgJ+B1wiqTm1P49wJqI6IiIYyraLAPOi4jlklokNaRjjAb+EHgihZwHgYtSX5cCd6f9XgfcBXwsIp4coMtiZnZEmNhSx/D6HNu6d51O2dKxnTGjGpg+1cv+bPcG/O0eyms9FqTHMgBz0s9vA9+WtBooAJem8NCniFgv6YvATyXtBJ4BPr6XY78R+AdJQfkRz40RsSrVXQe0Svp74CHS+hbgbyivWbm5/PSJomdLzMz2T22NmD2zidZFHWztKrGzZzvdO2teebunrm4wbkN2qNIecoH105RZs2LlwoV9rXPp5YtsZke0QqFIW76dDZs6Gd/cyPSpLQ4oh7c93RP7zb8hZmY24OrqapkxfcJQD8MOMf6z+GZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkm1Qz0AMzM7/BQKRdry7Tz3/Mu8ZtwIpk9toa7OtxzbN0P6GyPpauAqoAjcGxF/lcpPB74JjARKwJuBo4CfVTRvAf53RHxW0tuBbwCnA7Mj4vsVx7gB+CPKs0Y/Bv4iIqKifj7wiYhoTN8/B3w0VdcCbwSaI2Lz/p5nPp8nn8/vb3Mzs0PKxheLtC7qoKu7RH19AyNHNDJmVAPz553LpBPGDvXw7BAy4I97VPZ7x5H0TuB84PSIeBNwYyqvBf438OlU/g5gZ0S8HBFTej/AM8BdqbvfAh8H/r3qGG8D/pByeDmVctiZUVE/DRhV2SYivlpxjDnA0gMJKGZmR5JiT9C6qIMdO0uMHJ5j3JgGjm1upHNbgavn3kehUBzqIdohZEBCiqQJktZIuhlYASyWtFrSKknXpN2uAL4SETsAIuL5VP5e4JGIeDiVvxgRPVX9TwLGkWZWImJdRDxCedalUgD1QB1wNOXZmI2pjxrgq8Bf7eFULgG+s88XwMzsCLW2vUBXd4lh9bveXkY3NbD5pe205duHaGR2KBrImZSTgduBy4BiRJwaEacBC1L9ScDZkn4haamkN1eUh6QHJK2Q1FeIuAS4o/KxTV8iog14EHgufR6IiDWp+irgnoh4rq+2koYBM4E7d1N/uaTlkpbvaQxmZkeSjpdLlKr/dzHp6SmxYVPn4A7IDmkDGVKeiYhlwFpgoqT5kmYCW1N9LTAaeCvwOeC7kpTK/xvldSH/DbhA0ruq+p5NP2Y4JJ1IeU1JC3AccI6kt0t6LXAxMH8PzT8I/NfuHvVExC0RMS0ipu1tHGZmR4qmETlyu7mz1NTkGN/cOLgDskPaQIaULoCI2AJMBpYAVwK3pvp24K4o+yXlRzXHpPKlEfFCRGwD7gPO7O1U0mSgNiL6sxL1AmBZRHRGRCdwP+VQdAZwIvC0pHXAMElPV7XtVxAyM7NXTWypY3h9jm3du06nbOnYzphRDUyf2jJEI7ND0WAsnD0GyEXEncBcXg0cC4Fz0j4nUV438gLwAHC6pGFpEe0M4LGKLvdlnchvgRmSaiUdlfpaExH3RsT4iJgQEROAbRFxYsWYm9K+d+/XSZuZHaFqa8TsmU0cfVSOrV0lNm3ezsZNnTQOq2P+vHP9GrLtk8H4bTkOWFDxhs+c9PPbwLclrQYKwKVpjckWSV8DfkV54et9EXFvRX8fBs6tPEBaz/IDyo+PPijpi+nNoO9TDkKrUl+LIuKH/RjzBcCPIqJr30/3902dOpWpU6cejK7MzA4J1/1l+e+kbNjUyfjmRv+dFNsv2svaU+uHKbNmxcqFC7WHXXyRzczsSLKne2K/+c/im5mZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSYNWUiR9FVJj0t6RNIPJI2qqJsj6WlJT0h6XyobIWllxecFSd+o6vMiSSFpWlX5SEm/k3RTRdk5klZIWi3pNkm1qfwNktok7ZB07cBeBTOz7CkUiixtW0fr3atY2raOQqE41EOyI1TtQB9AkgBFRKmq6sfAnIgoSroemANcJ+kUYDbwJuC1wE8knRQRLwNTKvrNA3dVfB8BfAb4RR/DmAcsrdg3B9wGvCsinpT0d8ClwLeAzamfWQd25q/K5/Pk8/mD1Z2Z2YDZ+GKR1kUddHWXqK9vYOSIRsaMamD+vHOZdMLYoR6eHWEGZCZF0gRJayTdDKwAFqcZi1WSrgGIiB9FRG88Xwa0pO3zgdaI2BERvwGeBs6q6n8SMA74WUXxPOAGoLtq36nAscCPKorHAjsi4sn0/cfAhWlcz0fEr4Cd+38FzMwOPcWeoHVRBzt2lhg5PMe4MQ0c29xI57YCV8+9zzMqNugG8nHPycDtwGVAMSJOjYjTgAV97PtJ4P60fRzwbEVdeyqrdAlwR0QEgKQzgOMj4j8qd0ozJv8AfK6q/QvAURWPhS4Cjt+HczMzO+ysbS/Q1V1iWP2ut4bRTQ1sfmk7bfn2IRqZHakGMqQ8ExHLgLXAREnzJc0EtlbuJOnzQBH4t96iPvqKqu+zge+k9jng68Bf9tHuz4H7IqIy9JDCzWzg65J+CbycxtBvki6XtFzS8n1pZ2aWVR0vlyhVP5hPenpKbNjUObgDsiPeQIaULoCI2AJMBpYAVwK39u4g6VLgA8BHe2dFKM+cVM5qtADrK9pMBmojoneRxwjgVGCJpHXAW4F70izJdOCqVH4j8CeSvpLG1RYRZ0fEWcBPgaf25eQi4paImBYR0/a+t5lZ9jWNyJHbzV2hpibH+ObGwR2QHfEGY+HsMUAhIu6U9GvgX1P5TOA6YEZEbKtocg/w75K+Rnnh7CTglxX1l5BmUQAiogM4puJ4S4BrI2I58NGK8o8D0yLi/03fx0XE85KOTuP40sE6ZzOzQ9HEljqG1+fYVvXIZ0vHdsaMamD61JY9tDY7+AY8pFBeT7IgPZaB8ls8ADcBRwM/Lr8AxLKI+HREPCrpu8BjlB/BXBkRPRX9fRg49yCM63OSPkB5NumfIuL/AEgaDywHRgIlSZ8FTomIrbvvyszs0FdbI2bPbKJ1UQdbu0rs7NlO986aV97uqasbjFuG2av06lMW219TZs2KlQsX9rWWppcvspkdMgqFIm35djZs6mR8cyPTp7Y4oNi+2tM9sd/8W2dmZruoq6tlxvQJQz0MM/9ZfDMzM8smhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzy6SDElIkXSzpUUklSdP6qH+dpE5J11aUfUnSs5I6q/Z9vaTFkh6RtERSS0V5XtLKdKxPV7T5V0m/SXUrJU1J5U2Sfijp4dTmE6n85Ip9V0raKumzqe6OivJ1klYejGtkZnaoKBSKLG1bR+vdq1jato5CoTjUQ7IjVO2+7CxJgCKiVFW1GvgQ8M3dNP06cH9V2Q+Bm4CnqspvBG6PiNsknQN8GfgY8BzwtojYIakRWC3pnohYn9p9LiK+X9XXlcBjEfFBSc3AE5L+LSKeAHqDTA3wO+AHABHxkYrz/QegY3fXo7/y+Tz5fP5AuzEzG3AbXyzSuqiDru4S9fUNjBzRyJhRDcyfdy6TThg71MOzI8xeZ1IkTZC0RtLNwApgsaTVklZJugYgItakG39f7WcBa4FHK8sjYllEPNdHk1OAxWn7QeD8tH8hInak8qP7M3YggBEpXDUCm4Hq/yV4F/DriHimatwCPgx8px/HMTM75BV7gtZFHezYWWLk8BzjxjRwbHMjndsKXD33Ps+o2KDr7+Oek4HbgcuAYkScGhGnAQv21EjScOA64Iv7MKaHgQvT9gWUQ8bY1N/xkh4BngWur5hFAfhSekT0dUlHp7KbgDcC64FVwF/0MQs0m76DyNnAxoionukxMzssrW0v0NVdYlj9rreG0U0NbH5pO2359iEamR2p+htSnomIZZRnRCZKmi9pJrB1L+2+CHw9Ijr3sl+la4EZkh4CZlB+FFMEiIhnI+J04ETgUknHpjZzgDcAbwbGUA5GAO8DVgKvpfx45yZJI3sPJKkOOA/4Xh/juIQ9zKJIulzScknL9+HczMwyq+PlEqXq/41LenpKbNi0L/+Umx24/oaULoCI2AJMBpZQXu9x617avQW4QdI64LPAX0u6ak8NImJ9RHwoIs4APp/KOqr3ofz46Oz0/bko20F5duestOsngLtS3dPAbyiHmV7vB1ZExMbK/iXVUl5jc8cexnlLREyLiN9bKGxmdihqGpEjt5u7Qk1NjvHNjYM7IDvi7dPbPZKOAXIRcScwFzhzT/tHxNkRMSEiJgDfAP5nRNy0t2NI6h3XHODbqbxFUkPaHg38IfBE+v6a9FPALMoLeQF+S3nNCWnW5WTKs0G9djdb8m7g8Yjw3KaZHTEmttQxvD7Htu5dp1O2dGxnzKgGpk9tGaKR2ZFqX19BPg5Ykl7L/VfKIQJJF0hqB6YD90p6YG8dSbohtRkmqV3S36aqd1B+C+dJ4FjgS6n8jcAvJD0MLAVujIhVqe7fJK2ivO7kGODvU/k84G2pbjFwXUS8kI4/DHgPcFcfw9vdOhUzs8NWbY2YPbOJo4/KsbWrxKbN29m4qZPGYXXMn3cudXX79EKo2QFTRAz1GA55U2bNipULF2oPu/gim9kho1Ao0pZvZ8OmTsY3NzJ9aosDiu2rPd0T+82/dWZmtou6ulpmTJ8w1MMw85/FNzMzs2xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTHJIMTMzs0xySDEzM7NMckgxMzOzTBqykCJpsqQ2Sask/VDSyFT+Hkn5VJ6XdE4qHyFpZcXnBUnfSHVvl7RCUlHSRVXHeZ2kH0laI+kxSRNSuSR9SdKTqe4zqfwNaVw7JF07mNfEzGyoFQpFlrato/XuVSxtW0ehUBzqIdkRrHagDyBJgCKiVFV1K3BtRCyV9Engc8Bc4AXggxGxXtKpwAPAcRHxMjClot88cFf6+lvg40BfoeJ24EsR8WNJjUDvOD4OHA+8ISJKksal8s3AZ4BZB3Dau8jn8+Tz+YPVnZnZgNj4YpHWRR10dZeor29g5IhGxoxqYP68c5l0wtihHp4dgQZkJkXShDQ7cTOwAlgsaXWaHbkm7XYy8NO0/WPgQoCIeCgi1qfyR4F6SUdX9T8JGAf8LLVZFxGP8GoA6d3vFKA2In6c9uuMiG2p+grg73rDU0Q83/szIn4F7DwoF8PM7BBQ7AlaF3WwY2eJkcNzjBvTwLHNjXRuK3D13Ps8o2JDYiAf95xMeRbjMqAYEadGxGnAglS/GjgvbV9MeVaj2oXAQxGxo6r8EuCOiIi9jOEk4CVJd0l6SNJXJdWkuj8APiJpuaT7U/AxMzsirW0v0NVdYlj9rreF0U0NbH5pO2359iEamR3JBjKkPBMRy4C1wERJ8yXNBLam+k8CV6bHNiOAQmVjSW8Crgf+rI++ZwPf6ccYaoGzKT8GejMwkfJjHoCjge6ImAb8C/Dt/p8aSLo8BZzl+9LOzCyLOl4uUap+KJ/09JTYsKlzcAdkxsCGlC6AiNgCTAaWAFdSXotCRDweEe+NiKmUA8evextKagF+APxJRPy6slNJkyk/wunPIo92yjMxayOiCCwEzqyouzNt/wA4fV9OLiJuiYhpKeSYmR3SmkbkyO3mjlBTk2N8c+PgDsiMQXi7R9IxQC4i7qS8MPbMVD4u/cwBXwD+OX0fBdwLzImI/+qjy0vo3ywKwK+A0ZKa0/dzgMfS9sL0HWAG8OQ+nJaZ2WFlYksdw+tzbOvedTplS8d2xoxqYPrUliEamR3JBuMV5OOAJZJWAv8KzEnll0h6EngcWM+ra1WuAk4E5la8bjyuor8PUxVSJL1ZUjvltS3flPQoQET0UH7Us1jSKkCUH+0AfAW4MJV/mfLaGSSNT339d+ALktp7X482Mztc1daI2TObOPqoHFu7SmzavJ2NmzppHFbH/HnnUlc34C+Dmv0e7X3tqe3NlFmzYuXChdrDLr7IZnZIKBSKtOXb2bCpk/HNjUyf2uKAYvtjT/fEfvNvnpmZvaKurpYZ0ycM9TDMAP9ZfDMzM8sohxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxQzMzPLJIcUMzMzyySHFDMzM8skhxV/M0wAACAASURBVBQzMzPLJIcUMzMzy6TaoR6AmZkNrUKhSFu+neeef5nXjBvB9Kkt1NX59mBDb0B/CyVdDPwt8EbgrIhYXlE3B/gU0AN8JiIeSOWXAH8NBLAe+OOIeKGi3UXA94A39/Yn6VLgC2mXv4+I26rGMR/4REQ0pu9vABYAZwKfj4gbK/b9C+BPAQH/EhHfONDrkM/nyefzB9qNmdlBt/HFIq2LOujqLlFf38DIEY2MGdXA/HnnMumEsUM9PDvCHZTHPSrrq6/VwIeAn1btfwowG3gTMBO4WVKNpFrgH4F3RsTpwCPAVRXtRgCfAX5RUTYG+B/AW4CzgP8haXRF/TRgVNW4Nqd+bqwslHQq5YByFjAZ+ICkSf28DGZmh5RiT9C6qIMdO0uMHJ5j3JgGjm1upHNbgavn3kehUBzqIdoRbr9DiqQJktZIuhlYASyWtFrSKknXAETEmoh4oo/m5wOtEbEjIn4DPE05GCh9hksSMJLybEqvecANQHdF2fuAH0fE5ojYAvyYcvBBUg3wVeCvKg8eEc9HxK+AnVXjeiOwLCK2RUQRWApcsI+XxszskLC2vUBXd4lh9bveCkY3NbD5pe205duHaGRmZQc6k3IycDtwGVCMiFMj4jTKj1L25Djg2Yrv7cBxEbETuAJYRTmcnAJ8C0DSGcDxEfEf/ekrbV8F3BMRz/XzfFYDb5c0VtIw4Fzg+L52lHS5pOWSlvdVb2aWdR0vlyiV+q7r6SmxYVPn4A7IrMqBhpRnImIZsBaYKGm+pJnA1r20Ux9lIekoyiHlDOC1lB/3zEmPkr4O/OU+9PVa4GJgfv9OpTzzA1xPeTZmEfAw0Od8Z0TcEhHTImJaf/s3M8uSphE5cru5C9TU5Bjf3Di4AzKrcqAhpQsgPWaZDCwBrgRu3Uu7dnadoWihPHMyJfX364gI4LvA24ARwKnAEknrgLcC96T1Jrvr6wzgRODp1GaYpKf3dkIR8a2IODMi3k557cpTe2tjZnYomthSx/D6HNu6d51O2dKxnTGjGpg+tWWIRmZWdrAWzh4D5CLiTmAu5bdm9uQeYLakoyWdAEwCfgn8DjhFUnPa7z3AmojoiIhjImJCREwAlgHnpbd7HgDeK2l0WjD7XuCBiLg3IsZXtNkWESf241zGpZ+vo7zo9zv7ci3MzA4VtTVi9swmjj4qx9auEps2b2fjpk4ah9Uxf965fg3ZhtzB+g08DlhQ8YbPHABJF1B+3NIM3CtpZUS8LyIelfRd4DHKj1OujIgeYL2kLwI/lbQTeAb4+J4OHBGbJc0DfpWK/i4iNu+pjaTxwHLKC3NLkj4LnBIRW4E7JY2lvKj2yjRLdECmTp3K1KlTD7QbM7MBcd1flv9OyoZNnYxvbvTfSbHMUPmpih2IKbNmxcqFC/taG9PLF9nMzI4ke7on9pv/LL6ZmZllkkOKmZmZZZJDipmZmWWSQ4qZmZllkkOKmZmZZZJDipmZmWWSQ4qZmZllkkOKmZmZZZJDipmZmWWSQ4qZmZllkkOKmZmZZZJDipmZmWWSQ4qZmZllkkOKmZmZZZJDipmZmWWSQ4qZmZllkkOKmZmZZZJDipmZmWWSQ4qZmZll0pCFFElTJC2TtFLScklnVdTNkfS0pCckva+i/COSHpH0qKQbKspfL2lxqlsiqaWi7oa0/xpJ/0uSqsYxX1Jnxfd3SOpI41op6W8G7iqYmQ2tQqHI0rZ1tN69iqVt6ygUikM9JLNX1A70AVIoUESUqqpuAL4YEfdLOjd9f4ekU4DZwJuA1wI/kXQSMAr4KjA1IjZJuk3SuyJiMXAjcHtE3CbpHODLwMckvQ34Q+D0dMz/BGYAS9LYpqV+q/0sIj5wsK5BPp8nn88frO7MzA6KjS8WaV3UQVd3ifr6BkaOaGTMqAbmzzuXSSeMHerhmQ3MTIqkCWnm4mZgBbBY0mpJqyRdk3YLYGTabgLWp+3zgdaI2BERvwGeBs4CJgJPRsSmtN9PgAvT9inA4rT9YOqj9xj1QB1wNHAUsDGNsYZy6Pmrg3fmZmaHhmJP0Lqogx07S4wcnmPcmAaObW6kc1uBq+fe5xkVy4SBfNxzMnA7cBlQjIhTI+I0YEGq/yzwVUnPUp4JmZPKjwOereinPZU9DbwhBaBaYBZwfNrnYV4NLBcAIySNjYg2yqHlufR5ICLWpP2uAu6JiOf6GPt0SQ9Lul/Smw7gGpiZZdLa9gJd3SWG1e96Gxjd1MDml7bTlm8fopGZvWogQ8ozEbEMWAtMTGs/ZgJbU/0VwDURcTxwDfCtVK7f74qIiC2pzR3Az4B1QG/UvxaYIekhyo9zfgcUJZ0IvBFooRx0zpH0dkmvBS4G5vdxrBXA6yNicqpf2NfJSbo8raVZ3r/LYWaWHR0vlyhVP4RPenpKbNjU2Xel2SAayJDSBZDCxWTK60CuBG5N9ZcCd6Xt71F+pAPlmZPjeVUL6VFQRPwwIt4SEdOBJ4CnUvn6iPhQRJwBfD6VdVCeVVkWEZ0R0QncD7wVOAM4EXha0jpgmKSnU7utaV8i4j7gKEnHVJ9cRNwSEdMiYtr+XyIzs6HRNCJHbjd3gJqaHOObGwd3QGZ9GPC3e9INPhcRdwJzgTNT1XrKsx4A55ACB3APMFvS0ZJOACYBv0x9jUs/RwN/Tgo8ko6R1Hsuc4Bvp+3fUp5hqZV0VDremoi4NyLGR8SEiJgAbIuIE1Nf43vfAEpvHOWAFw/qRTEzG2ITW+oYXp9jW/eu0ylbOrYzZlQD06e27Kal2eAZ8Ld7KD9mWVAVIgD+FPjHtL6kG7gcICIelfRd4DHKj3OujIie1OYfJU1O238XEU+m7XcAX5YUwE8pz9gAfJ9yAFpFeRHtooj44V7GexFwhaQisB2YHRGxH+dtZpZZtTVi9swmWhd1sLWrxM6e7XTvrHnl7Z66usG4PZjtmXz/PXBTZs2KlQsX9rWWppcvspllUqFQpC3fzoZNnYxvbmT61BYHFDsY9nRP7Df/JpqZHcHq6mqZMX3CUA/DrE/+s/hmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJQxZSJE2W1CZplaQfShqZyt8jKZ/K85LOqWhTJ+kWSU9KelzShan8aEl3SHpa0i8kTaho0yNpZfrcU1H+r5J+U1E3JZW/Q1JHRfnfDNY1MTMbaIVCkaVt62i9exVL29ZRKBSHekhmu1U70AeQJEARUaqquhW4NiKWSvok8DlgLvAC8MGIWC/pVOAB4LjU5vPA8xFxkqQcMCaVfwrYEhEnSpoNXA98JNVtj4gpuxne5yLi+32U/ywiPrAfp9unfD5PPp8/WN2Zme2XjS8WaV3UQVd3ifr6BkaOaGTMqAbmzzuXSSeMHerhmf2eAZlJkTRB0hpJNwMrgMWSVqfZkWvSbicDP03bPwYuBIiIhyJifSp/FKiXdHT6/kngy2m/UkS8kMrPB25L298H3pXCkZmZAcWeoHVRBzt2lhg5PMe4MQ0c29xI57YCV8+9zzMqlkkD+bjnZOB24DKgGBGnRsRpwIJUvxo4L21fDBzfRx8XAg9FxA5Jo1LZPEkrJH1P0rGp7DjgWYCIKAIdQO//FtRLWi5pmaRZVf1/SdIjkr5eEYQApkt6WNL9kt60f6dvZpYda9sLdHWXGFa/6z/7o5sa2PzSdtry7UM0MrPdG8iQ8kxELAPWAhMlzZc0E9ia6j8JXCkpD4wACpWNUzi4HvizVFQLtAD/FRFnAm3Ajb2793H8SD9fFxHTgP8H+IakP0jlc4A3AG+m/NjoulS+Anh9REwG5gML+zo5SZen8LN875fCzGxodbxcolT90D3p6SmxYVPn4A7IrB8GMqR0AUTEFmAysAS4kvJaFCLi8Yh4b0RMBb4D/Lq3oaQW4AfAn0REb/mLwLZUDvA94My03U6aiZFUCzQBm9Nx1qefa9MYzkjfn4uyHZRnd85K5VsjojNt3wccJemY6pOLiFsiYloKQGZmmdY0IkduN//i19TkGN/cOLgDMuuHAX+7J93gcxFxJ+WFsWem8nHpZw74AvDP6fso4F5gTkT8V28/ERHAD4F3pKJ3AY+l7XuAS9P2RcD/iYiQNLr3MU4axx/2tpH0mvRTwCzKj5+QNL53PYuksyhfoxcP3hUxMxt8E1vqGF6fY1v3rtMpWzq2M2ZUA9OntgzRyMx2b8Df7qG8XmRBCiNQfswCcImkK9P2Xby6VuUq4ERgrqS5qey9EfE85Ucy/7+kbwCbgE+k+m+l8qcpz6DMTuVvBL4pqUQ5bHwlInqDzb9Jaqb8qGgl8OlUfhFwhaQisB2YnQKSmdkhq7ZGzJ7ZROuiDrZ2ldjZs53unTWvvN1TVzcYtwOzfSPffw/clFmzYuXChXt6m8gX2cwyoVAo0pZvZ8OmTsY3NzJ9aosDig2Eg/KGrX8zzcyOIHV1tcyYPmGoh2HWL/6z+GZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkOKWZmZpZJDilmZmaWSQ4pZmZmlkkHJaRIuljSo5JKkqZVlI+V9KCkTkk3VbVZIukJSSvTZ1wq/++SHpP0iKTFkl5f0eZ6SavT5yMV5e+StCL185+STkzlb5DUJmmHpGsr9j+54rgrJW2V9NlUd0dF+TpJKw/GNTIzG2qFQpGlbetovXsVS9vWUSgUh3pIZntUuy87SxKgiChVVa0GPgR8s6q8G5gLnJo+1T4aEcuryh4CpkXENklXADcAH5H0R8CZwBTgaGCppPsjYivwT8D5EbFG0p8DXwA+DmwGPgPMqjxARDyR+kFSDfA74AeprjL8/APQsceL0g/5fJ58Pn+g3ZiZ7beNLxZpXdRBV3eJ+voGRo5oZMyoBubPO5dJJ4wd6uGZ9WmvMymSJkhaI+lmYAWwOM1krJJ0DUBErEk3/l1ERFdE/CflsNIvEfFgRGxLX5cBLWn7FGBpRBQjogt4GJjZ2wwYmbabgPWpr+cj4lfAzj0c8l3AryPimarzFvBh4Dv9HbuZWRYVe4LWRR3s2Fli5PAc48Y0cGxzI53bClw99z7PqFhm9fdxz8nA7cBlQDEiTo2I04AFB3j8BemxytwUCqp9Crg/bT8MvF/SMEnHAO8Ejk91lwH3SWoHPgZ8ZR/GMJu+g8jZwMaIeGof+jIzy5y17QW6uksMq9/1n/zRTQ1sfmk7bfn2IRqZ2Z71N6Q8ExHLgLXAREnzJc0Eth7AsT+ags7Z6fOxykpJfwxMA74KEBE/Au4Dfk45VLQBvfH/GuDciGihHJy+1p8BSKoDzgO+10f1JexhFkXS5ZKWS6p+XGVmlikdL5coVT+kT3p6SmzY1Dm4AzLrp/6GlC6AiNgCTAaWAFcCt+7vgSPid+nny8C/A2f11kl6N/B54LyI2FHR5ksRMSUi3gMIeEpSMzA5In6RdrsDeFs/h/F+YEVEbKwslFRLeY3NHXsY/y0RMS0ipu1uHzOzLGgakSO3m3/ta2pyjG9uHNwBmfXTPr3dkx6z5CLiTsoLYs/cn4NKqk19Ieko4AOUF98i6QzKC3DPi4jnK9rUSBqbtk8HTgd+BGwBmiSdlHZ9D7Cmn0PZ3WzJu4HHI8JzoGZ2yJvYUsfw+hzbunedTtnSsZ0xoxqYPrVlNy3NhtY+vd0DHEd5HUlvuJkDIOkCYD7QDNwraWVEvC/VraO8qLVO0izgvcAzwAMpoNQAPwH+JfX5VaAR+F5apvLbiDgPOAr4WSrbCvxxRBTTMf4UuFNSiXJo+WQqHw8sT8cvpdeMT4mIrZKGUQ40f9bHee5unYqZ2SGntkbMntlE66IOtnaV2Nmzne6dNa+83VNXt6+3ArPBoYgY6jEc8qbMmhUrFy7sa+FvL19kMxtyhUKRtnw7GzZ1Mr65kelTWxxQbKDs6Z7Yb/7tNDM7QtTV1TJj+oShHoZZv/nP4puZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSY5pJiZmVkmOaSYmZlZJjmkmJmZWSbVDvUAzMxs/xQKRdry7Tz3/Mu8ZtwIpk9toa7O/6zb4WPIfpsl3QGcnL6OAl6KiCmSPgp8rmLX04EzI2KlpDrgJuAdQAn4fETcKem/A5cBRWAT8MmIeCYd53rgj1Jf8yLijlR+AtAKjAFWAB+LiIKk0cC3gT8AulNfqw/kXPP5PPl8/kC6MDPbxcYXi7Qu6qCru0R9fQMjRzQyZlQD8+edy6QTxg718MwOigF/3KOy3ztORHwkIqZExBTgTuCuVP5vFeUfA9ZFxMrU7PPA8xFxEnAKsDSVPwRMi4jTge8DN6Rj/xFwJjAFeAvwOUkjU5vrga9HxCRgC/CpVP7XwMrU158A/3iwroWZ2cFQ7AlaF3WwY2eJkcNzjBvTwLHNjXRuK3D13PsoFIpDPUSzg2JAQoqkCZLWSLqZ8izFYkmrJa2SdE3VvgI+DHynj64uqSr/JPBlgIgoRcQLafvBiNiW9lkGtKTtU4ClEVGMiC7gYWBmOuY5lAMNwG3ArIo2i1O/jwMTJB27XxfCzGwArG0v0NVdYlj9rv+Ej25qYPNL22nLtw/RyMwOroGcSTkZuJ30GCYiTo2I04AFVfudDWyMiKf66OMjpJAiaVQqmydphaTv7SY8fAq4P20/DLxf0jBJxwDvBI4HxlJ+vNT7vxvtwHEVbT6UjnkW8HpeDT2vkHS5pOWSlu/xKpiZHWQdL5colfqu6+kpsWFT5+AOyGyADGRIeSYilgFrgYmS5kuaCWyt2q96tgQASW8BtlWsB6mlHBb+KyLOBNqAG6va/DEwDfgqQET8CLgP+Hk6RhvldSvqY7yRfn4FGC1pJXA15UdJvzd3GhG3RMS0iJi2x6tgZnaQNY3IkdvNv941NTnGNzcO7oDMBshAhpQugIjYAkwGlgBXArf27iCplvKsxR19tJ/NruHlRWAb8IP0/XuU15v09vVuymtWzouIHb3lEfGltMblPZTDyVPAC8CodHwoh5/1af+tEfGJtCbmT4Bm4Df7cf5mZgNiYksdw+tzbOvedTplS8d2xoxqYPrU35v8NTskDcbC2WOAXETcCcylIlgA7wYej4j2qjY54GLKb98AEBEB/JDymz0A7wIeS/ufAXyTckB5vqKfGklj0/bplN8U+lHq60HgorTrpcDdab9R6S0iKD+q+mlEVM/+mJkNmdoaMXtmE0cflWNrV4lNm7ezcVMnjcP+b3v3HiZXUed//P1NJkMSJleSGJawJDEQIIEE0mbJqtCgSMRWsutlk8cVlUXWW1hRkGURcQVXlHVBgsoid0XyU8TAdiDCg3RAHYSeEIQAQojBDBCTALlfJjPz/f1R1eTQdCbD9EzPIXxezzNPn65Tp06d6p4+366qc7qeeRedrMuQZa9Ri3fyAcD1iSt8zkusK+8tKTkWaHb3FWXp5wI/MbPLCZcafzqmXwo0AL8Ic2L5i7t/COgHPBDTNgL/nJiHci4w38wuJgzpXBvTDwNuMrM2QhBUuuqny6ZNm8a0adOqLUZE5DXO/Uq4T8rqtZsZPbJB90mRvY6FTgWpxtRZs3zpggWV5rmUqJFFROStpKNzYqfptvgiIiKSSgpSREREJJUUpIiIiEgqKUgRERGRVFKQIiIiIqmkIEVERERSSUGKiIiIpJKCFBEREUklBSkiIiKSSgpSREREJJUUpIiIiEgqKUgRERGRVFKQIiIiIqmkIEVERERSSUGKiIiIpJKCFBEREUklBSkiIiKSSgpSREREJJUUpIiIiEgq1SRIMbOzzczNbER8vp+Z3Wdmm83sykS+gWa20MyeMrNlZnZJYt1lZrY0/j1tZusT6xaZ2Xozy5ft9wQzW2Jmj5vZjWZWF9OzZrYhUd7XE9tcZ2ZrzOzxnmwTEZFqtLS0srhxJfNvf4zFjStpaWnt7SqJdLu67ijEzAwwd2+vsO5A4ETgL4nk7cAFwOT4l/Tf7n6fmdUD95rZ+939Lnc/K1HmXOCoxDaXAgOBf03k6QPcCLzH3Z82s28CnwSujVkecPdchcO5AbgSuGnPR945TU1NNDU1dVdxIvIW99eXWpm/aANbtrfTv/8ABg9qYPjQAcy76GQOHrdfb1dPpNt0uSfFzMaa2ZNm9kNgCSGgeNzMHjOzsxJZLwO+Cngpwd23uPtvCcEKifSt7n5fXG6J5Y6psPs5wC2J7e4FNpXl2Q/Y4e5Px+f3AB/e03G5+/3Ay3vKJyLSG1rbnPmLNrBjZzuD9+3DqOEDeNvIBjZvbWHuBXeqR0X2KtUO90wk9DicDrS6+2R3PwK4HsDMPgQ87+6PvtGCzWwo8EHg3rL0g4BxwG/2UMQ6oJ+ZZeLzjwAHJtbPMLNHzewuM5v0RusnItIbVjS3sGV7OwP7v/bje9iQAby8fhuNTc29VDOR7ldtkPKcuz8IrADGm9k8M5sJbDSzgcD5wNc7LKGCOHfkFuAKd19Rtno2cKu7t3VUhrt7zHuZmT1E6GkpfcVYAhzk7lOAecCCLtTxDDMrmlnxjW4rItJVGza10/66gfWgra2d1Ws317ZCIj2o2iBlC4C7vwJMAQrAF4BrgLcTejweNbOVhGGbJWY2uhPlXg084+6XV1g3m8RQT0fcvdHd3+3u04H7gWdi+kZ33xyX7yT0uIzoTJmJsq9294y7Z/acW0SkewwZ1Ic+u/nk7tu3D6NHNtS2QiI9qLsmzo4AWtz9l2b2LHCDuz8GjErkWQlk3H3dHsq6GBhCGEIqXzcRGAY0drJeo9x9jZntA5wLfCumjwb+6u5uZtMJwdpLnSlTRKQ3jR9Tz779+7C1bMjnlQ3bGD50ADOmVZrGJ/Lm1C1BCnAAcH28ogbgvD1tEIOWwUC9mc0C3gdsJAwRPUXodQG40t2viZvNAebHoZxkWQ8AhwINZtYM/Iu7/xo4x8xyhCDkR+5emsfyEeBzZtYKbANml8o0s1uALDAilnWhu1+LiEgK1PU1Zs8cwvxFG9i4pZ2dbdvYvrPvq1f31Nd318e6SO+zsvO9dMHUWbN86YIF1kEWNbKIdKuWllYam5pZvXYzo0c2MGPaGAUokiYdnRM7Te9oEZE3ofr6Oo6bMba3qyHSo3RbfBEREUklBSkiIiKSSgpSREREJJUUpIiIiEgqKUgRERGRVFKQIiIiIqmkIEVERERSSUGKiIiIpJKCFBEREUklBSkiIiKSSgpSREREJJUUpIiIiEgqKUgRERGRVFKQIiIiIqmkIEVERERSSUGKiIiIpJKCFBEREUklBSkiIiKSSj0epJjZXDP7k5ktM7PvxrQTzazJzB6Ljyck8i8ys0dj/qvMrG9Mv8zMlsa/p81sfWKbvzWzu83sSTN7wszGltVhnpltTjzPmtmGRHlfT6ybGeu73Mz+vedaRkSk81paWlncuJL5tz/G4saVtLS09naVRHpcXXcUYmYGmLu3l6UfD5wCHOnuO8xsVFy1Dvigu79gZpOBXwMHxHUfc/eNscxbgY8C8939rES5c4GjEru6CfiWu99jZg1AeyJvBhhaodoPuHuurL59gR8AJwLNwMNmdoe7P/GGGqRMU1MTTU1N1RQhIm9hf32plfmLNrBlezv9+w9g8KAGhg8dwLyLTubgcfv1dvVEekyXe1LMbGzsufghsAS418wej70jpYDic8Al7r4DwN3XxMdH3P2FmGcZ0N/M9onrNsb0OqAe8Aq7nwPcEutxOFDn7vfE7Te7+9a4ri9wKfDVTh7WdGC5u69w9xZgPiHIEhHpFa1tzvxFG9ixs53B+/Zh1PABvG1kA5u3tjD3gjvVoyJ7tWqHeyYSejFOB1rdfbK7HwFcH9cfArzbzP5gZovN7B0Vyvgw8EgpkAEws18Da4BNhN4UEusOAsYBv0nsY72Z3WZmj5jZpaUhIuCLwB3u/mKF/c6Iw0p3mdmkmHYAsCqRp5ldPTwiIjW3ormFLdvbGdj/tR/Xw4YM4OX122hsau6lmon0vGqDlOfc/UFgBTA+zv2YCSR7Q4YBxwDnAD+PwzgAxODgO8C/Jgt195OA/YF9gBN4rdnAre7eltjHu4GzgXcA44FPmdnfEIaK5lWo9xLgIHefEtcvKFWpQt5KPTmY2RlmVjSzYqX1IiLdYcOmdtrbK69ra2tn9drNlVeK7AWqDVK2ALj7K8AUoAB8Abgmrm8GbvPgIcJckREAZjYG+BVwqrs/W16wu28H7uD1wy2ziUM9iX08EodoWgkBx9GEOSsTgOVmthIYaGbLY9kb3X1zXL4T6GdmI2JZBybKHgO8QAXufrW7Z9w902ELiYhUYcigPvTZzSd13759GD2yobYVEqmhbrm6J57g+7j7L4ELCEEChIDhhJjnEMIck3VmNhRYCJzn7r9LlNNgZvvH5TrgZOCpxPqJhJ6ZxsTuHwaGmdnI+PwE4Al3X+juo919rLuPBba6+4RYzuhSj46ZTY/t8FIs62AzG2dm9YSA6I7uaCMRka4YP6aeffv3Yev213anvLJhG8OHDmDGtDG9VDORntctV/cQ5m1cb2aloOe8+HgdcJ2ZPQ60AJ90dzezLxJ6OS4wswti3vcRhlvuiJNo+xLmnVyV2M8cwpU+rw7BuHubmZ1NmLhrQBPw4z3U9yPA58ysFdgGzI5ltsa6/Tru9KIzpAAAErhJREFU/zp3X/aGW0NEpJvU9TVmzxzC/EUb2LilnZ1t29i+s++rV/fU13fXx7hI+ljifC9dNHXWLF+6YEGl+SwlamQRqUpLSyuNTc2sXruZ0SMbmDFtjAIUSbOOzomdpne4iMibQH19HcfNGNvb1RCpKd0WX0RERFJJQYqIiIikkoIUERERSSUFKSIiIpJKClJEREQklRSkiIiISCopSBEREZFUUpAiIiIiqaQgRURERFJJQYqIiIikkoIUERERSSUFKSIiIpJKClJEREQklRSkiIiISCopSBEREZFUUpAiIiIiqaQgRURERFJJQYqIiIikUq8FKWZ2qZk9ZWZ/NLNfmdnQmF5vZteb2WNm9qiZZWP6QDNbGLdZZmaXJMq6zMyWxr+nzWx9Yt3fmtndZvakmT1hZmNj+hfNbLmZuZmNSOQ/1MwazWyHmZ1do+YQEdmtlpZWFjeuZP7tj7G4cSUtLa29XSWRmqjr6R2YmQHm7u1lq+4BznP3VjP7DnAecC7wGQB3P8LMRgF3mdk74jb/7e73mVk9cK+Zvd/d73L3sxL7mwscldjPTcC33P0eM2sASvX4HZAHCmX1ehk4E5hV3ZHv0tTURFNTU3cVJyJvIX99qZX5izawZXs7/fsPYPCgBoYPHcC8i07m4HH79Xb1RHpUj/SkmNnY2HPxQ2AJIaB4PPaOnAXg7ne7e+nrwIPAmLh8OHBvzLMGWA9k3H2ru98X01tiuWN4vTnALbEehwN17n5P3G6zu2+Ny4+4+8ryjd19jbs/DOysuiFERKrQ2ubMX7SBHTvbGbxvH0YNH8DbRjaweWsLcy+4Uz0qstfryeGeiYRejNOBVnef7O5HANdXyHsacFdcfhQ4xczqzGwcMA04MJk5Dg19kBjMJNIPAsYBv4lJhwDrzew2M3skDjH17Z7DExHpWSuaW9iyvZ2B/V/7UT1syABeXr+NxqbmXqqZSG30ZJDynLs/CKwAxpvZPDObCWxMZjKz84FW4OaYdB3QDBSBy4Hfx/Wl/HWEnpIr3H1F2T5nA7e6e1t8Xge8GzgbeAcwHvhUdxycmZ1hZkUzK3ZHeSIi5TZsaqe9fKA8amtrZ/XazbWtkEiN9WSQsgXA3V8BphDmfnwBuKaUwcw+CeSAj7u7x/yt7n6Wu09191OAocAziXKvBp5x98sr7HM2cagnagYecfcVcWhpAXB0dxycu1/t7hl3z3RHeSIi5YYM6kOf3XxK9+3bh9EjG2pbIZEaq8XE2RFAi7v/0syeBW6I6TMJE2WPK80TiekDCRNtt5jZiYShoifiuouBIYQhpPL9TASGAY2J5IeBYWY20t3XAicQemhERFJv/Jh69u3fh61lQz6vbNjG8KEDmDGt0rQ8kb1HLS5BPgAomNlSQoByXky/EhgE3BMvHb4qpo8ClpjZk4Qg5hMAZjYGOJ8wsXZJ3CYZrMwB5pd6ZADisM/ZhIm7jwEG/DiWd6aZNRMm3/7RzK6J6aNj+peBr5lZs5kN7t4mERHZs7q+xuyZQ9inXx82bmln7cvb+OvazTQMrGfeRSdTX9/j3zNFepUlzunSRVNnzfKlCxZYB1nUyCLSZS0trTQ2NbN67WZGj2xgxrQxClAk7To6J3aa3uUiIilXX1/HcTPG9nY1RGpOt8UXERGRVFKQIiIiIqmkIEVERERSSUGKiIiIpJImztbA1Fnd9luFIiIibwZ7uuq1UxSk1EBdczPFou4htyeZTEbt1Elqq85RO3We2qpz1E6dY2ZN3VGOhntEREQklRSkiIiISCopSKmBM844o7er8Kagduo8tVXnqJ06T23VOWqnTru6OwrRbfG7gW6LLyIi8hrdclt89aSIiIhIKilIqYKZzTSzP8Xlf6+w3szsigkTJnDkkUeyZMmSV9ctWrSIiRMnMmHCBC655JIa1rp37Ol43Z0zzzyTSm112mmnMWrUKCZPnlzLKveKrrbTqlWrOP744znssMOYNGkS3//+92td9Zrraltt376d6dOnM2XKFCZNmsSFF15Y66rXVDX/ewBtbW0cddRR5HK5WlW5V1TTTmPHjuWII45g6tSpZDKZWla7V3SmrczsCjNbbmZ/NLOjS+vMbKiZ3WpmT5nZk2Y2o8Odubv+uvAH9AWeBcZPOeUUBx4FDi/LczJwV3t7uzc2Nvr06dPd3b21tdXHjx/vzz77rO/YscOPPPJIX7Zsme+tOnO8Cxcu9JkzZ3p5W7m7L1682JuamnzSpEm1rnpNVdNOL7zwgjc1Nbm7+8aNG/3ggw/We2o3bdXe3u6bNm1yd/eWlhafPn26NzY21vwYaqHa/z139+9973s+Z84c/8AHPlDLqtdUte100EEH+dq1a2td7V7R2bYC7iIM+RwD/MF3nRdvBE6Py/XAUO/gXKuelK6bDix39xXx+XzglLI8pwA3mRnHHHMM69ev58UXX+Shhx5iwoQJjB8/nvr6embPns3tt99e08rXUmeO9/bbb+fUU0+lvK0Ajj32WIYPH94bVa+patpp//335+ijw5eVQYMGcdhhh/H888/3xmHURDVtZWY0NDQAsHPnTnbu3IlZtwyfp061/3vNzc0sXLiQ008/vTeqXzPVttNbSWfbCrgpxiwPAkPNbH8zGwwcC1wL4O4t7r6+wx12FMHor8OelI8A17g7sSflE8CVZXnywLsSz+8FMsltY/rrtt2b/jpzvLtrq8TzscDjvX0saW+nRFv9BRjc28eU1rYi9IQuBTYD3+nt40lxO90KTAOyQL63jyfF7fRnYAnQBJzR28eT1rYCpgIPATcAjwDXAPt2tD/1pHTdq1+9Elf2lF/FU+nrmXeQvrfqzPG+1dqkkqrbycwagF8CX3L3jd1Yt7Spqq3cvc3dpwJjgOlmtrdOeOpyO5lZDljj7t1y59CUq/Z/753ufjTwfuALZnZsd1YuZappqzrgaOBH7n4UsAV43XzOJAUpXdcMHJh4PgZ4oZN5OrPt3qSatnorqaqdzKwfIUC52d1v68F6pkG3vKc8dDUXgJndX8VUqKad3gl8yMxWEoazTzCzn/ZcVXtVVe8ndy89rgF+RZgOsLeq9tzX7O5/iOm3EoKW3VKQ0nUPAweb2TgzqwdmA3eU5bkDODVe5XMMsMHdX+zktnuTatrqraTL7WRhUsW1wJPu/j+1rXavqKatRprZUAAzGwC8F3iqlpWvoS63k7uf5+5j3H1s3O437v7PNa197VTzftrXzAYBmNm+wPuAx2tZ+Rqr5j21GlhlZhNjvvcAT3S0M/3AYBe5e6uZfRH4NWF8+zp3X2Zmn43rrwLuJFzhsxzYCny6o2174TBqopq2AjCzWwhj4iPMrBm40N2vre1R9Lwq2+mdhLHhx8xsaUz7D3e/s5bHUCtVttX+wI1m1pfwRe3n7p6v9THUQrX/e28VVbbT24BfxcnXdcDP3H1RjQ+hZrrhPTUXuDkGOCvYw/tNd5wVERGRVNJwj4iIiKSSghQRERFJJQUpIiIikkoKUkRERCSVFKSIiIhIKukSZJEUyWRzNwCfLEveRLjvwrxiIX9LF8tdCRwELC4W8tk3uG2WcAk4wOXFQn59Yt03gNLPCI8rFvIru1K/rshkc6VLE58rFvJja7Xf7pTJ5sYCn4pPFxQL+aW7zy3y1qOeFJH0GwTMAH6Wyeb+qRf2nyUEIhcCQ3th/3uzsexq26m9WxWR9FFPikh6HQ80AmcB345pZwD/r9dqVKZYyH8D+EYvV+NNJ5PNlW4kJyIdUJAikmLFQn5HJpv7AbuClOTvYZDJ5vYBzgXmAOOA7cDvgK8XC/kOfxguk82dRrjb48HA8LjtY8BlxUL+1pinAByX2OzPmWwO4hBL+XAP4e6bD8bnpxUL+esT+3s67uv+YiF/XEz7O+B8wh1zBwErgZ8AlxQL+Z0dt07FY0rW5z3A14BjgD8BnyXcAfNHhLthvgh8s1jI/ySx/UrisBhwJfBNYDzwJPClYiG/OJF3APAfwMcIPSJbgT8AFxcL+d/upsxrgK/HMvPAKYnqX5/J5krtNY7w42tXEH6FeDTQH1gN3AWcXyzk18Xys8B9cbvPA4cCH4/P7wDmFgv5LYn6HB3r/W5gGLAG+G2xkJ+dyHMS8NW47/7AM8BVwA+LhbzuACo1o0heJP2Svyi6prSQyebqCCes/yScmPYBhhBOwL/LZHPv3EO57wPeRQgs+hGChL8HfpHJ5k7uSkWLhfwfCCc0gI8m6jqFEKAA/DSmnQQ8AHyQECT1i3m+CXTHDyT+gtAbNYAwlHIHsDDWa19gAnBjJps7vMK2RxJ6rA4jtOtUYFEmm3t7rHs/4G5CEHQIUE8YCjsJKGSyuVyFMqcAN8Vj7NuJ+u9H+F2UgwmvTT9CkHpGPI5Kvg2cGbfdjxCEXlBamcnmTiQEkR8GRsUyDwD+KZHnXwjvqxMI76d9gMmEoG1eJ+ot0m0UpIikWOwp+Xwi6dbE8hzCSRjCZNsBhG/gTxJOLN/bQ/FXE06cQwknq0OBVXHdZwHiJNv/TGwzrljI2x4mqt4cH9+byeZKc1g+Eh93EIIHgB/E/f6e0BMxgDC0BZDLZHPV/jLxw4QT9aXx+SjgbwjBSak+BvxjhW2HAV8hnKTPjGn9CT0QEHoq3hWXbyYEWccBmwkByLxMNlf+c/VDgR/GOo0FTmXX6wfw6di2FicgrwFmEYKIUgD6jZh3euwRKddK+AXecYReFxLHCqEXqR/QTniNhxJ6eS4AyGRzDcD/xHa5jfA7Rw3sei99PpPNHVZhvyI9QsM9Iul1X2K5lfAt9opE2vsTyzfGv6TpmWxuYLGQ37qb8lcDFxOGWvbjtd/uD+lSjYObCSfTfsCHCL0HH47rFhYL+fWZbO4Q4O0x7e8Jwzzljgeq+aG27xYL+Zcz2dxvgHNi2g3FQv7ZTDa3KpHvwArbrioW8pfH5XmZbO6cmK/UO3VSIu/5xUL+FeD+TDZ3K+FqnbGEHpCnE/leAb5SLOR3AC8DxKGz3XmF8Dp8M5Y1oGz9IcCSsrRri4X8w7Hs+wlDUQfG58k2/1WxkP/fuLyB8D6A8FoMjsv/yOsDOCNMpH6yo4qLdBf1pIi8ORjhG23SyE5sM6zSikw2N4TwK6b/QOhhKB9+6N+FOgJQLOSXE+ZmAHw0k81NJgybQBzqYc91h9A7UY2/xMft5WnFQr4lkbZPhW2by54/Hx8PiI8jdpM3uVx+jE/HAKWzzgK+Sxh6Kg9QoPJr9ExiuXTc9RXq86fd7LMWr4tIp6knRSS9jifcH+XHhG7/zxBOLqWu93XxsR0YEb/NvyqTzVkHkxyPAcbE5UuAi4qF/NZMNlckTJZM6spEyZuBvwNOZNeJ8xV2zaVYl8h7ebGQP4syFYZL3qjWTqZVMqbseSk4KQUr68rWlQKi5HbJPPDaYKmko7YtzelZTXgv/An4APB/HWyTPL7ystcmlifuZvtknb9ULOS/n1y5h/eUSLdTkCKSYsVCfl0mmzudcKXKIOBrmWzummIhv4EwFDKb0CN6VSab+zLhJHMoYcihDpi7m6LrE8tbAM9kcx8HKs1zSAY/k6k8NFNuPmFuwz6JOvwi0YPxNPBnwtyJ0zPZ3N2E4a1B8Vi/TBiq6My+esKBmWxuLnADYb5PaUjod/HxbkLbA1ycyebOJLRNaVjrOV471LM7ybY9PJPN9S0W8m3xeek1aiPc0O8AwpVcXVIs5J/OZHPLCXNy/iGTzX2GMDl4MPCJYiH/bcL8oE2E1+GcGLQWCcOBJxPm6WhOitSMhntEUq5YyL/ErqsqhhJO4AA/A+6Pyx8jDDVsB5YCXyKcaHbn90DpzrEXES6fvRZ4oULehxPL/5fJ5jyTzf20Qr5kndcC98SnpS9DP02sd+CLhG/+DcCdwDbCZNFbgHd0VH4NrAMuAzayq+23A/8Vl39KuIcNwCcIwcYDhDZvA/6tkz0Oy9kVqJwDtGayudKQUanX6QDCa7uKMJG1Gp8HdhI++68mzEdZRTyuYiG/iV3zdw4Afks47ucJPXqHVrl/kTdEQYrIm8P3CN9wAb6UyeaGxfuInES4L8iThCtnNhDudXI5oSejohj4fIgw8XI7sIxwz47lFfI2Eq5qWUUYWuqsZCDzHOGElyz3TsK9Ou4AXgJa4j7uIlxmWylgqpVlhEmjT8R6LQXeXyzknwWIbf9ewsl9OeHEv4HQw3JCsZC/vTM7iZOaPxH311K2+r8Il/2ui2VfC/xbNQdVLOTvIQz13UYY/tlJCEB+nsjzv4RJ2ffG/e4g9Hrdxq77r4jUhLlreFFEBKr7jSMR6X7qSREREZFUUpAiIiIiqaThHhEREUkl9aSIiIhIKilIERERkVRSkCIiIiKppCBFREREUklBioiIiKSSghQRERFJpf8PuC6UEZ81M7gAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "features = SNP_outcome_cov_table_training_x.columns.values.tolist()\n",
    "features=[x.rsplit(\"_\")[0] for x in features]\n",
    "importances = regressor_xgb.feature_importances_\n",
    "\n",
    "mapping_table=pd.concat([pd.DataFrame(features),pd.DataFrame(importances)],axis=1)\n",
    "mapping_table.columns=[\"ID\",\"importance\"]\n",
    "\n",
    "rsid_table=pd.read_csv(\"../tables/time-germline.csv\",sep=\"\\t\")\n",
    "rsid_table=rsid_table.drop_duplicates(subset=\"rsid\")\n",
    "rsid_table=rsid_table[[\"snp_noallele\",\"rsid\"]]\n",
    "\n",
    "compiled_icb_rsid_plot=pd.merge(mapping_table,rsid_table,left_on=\"ID\",right_on=\"snp_noallele\")\n",
    "\n",
    "compiled_icb_rsid_plot=compiled_icb_rsid_plot.sort_values(by=\"importance\")\n",
    "compiled_icb_rsid_plot=compiled_icb_rsid_plot.iloc[compiled_icb_rsid_plot.shape[0]-15:compiled_icb_rsid_plot.shape[0],:]\n",
    "\n",
    "plt.rcParams['font.family'] = 'sans-serif'\n",
    "plt.rcParams['font.sans-serif'] = 'Helvetica'\n",
    "\n",
    "plt.rcParams['axes.edgecolor']='darkslategrey'\n",
    "plt.rcParams['axes.linewidth']=1\n",
    "plt.rcParams['xtick.color']='black'\n",
    "plt.rcParams['ytick.color']='black'\n",
    "plt.style.use('seaborn-dark-palette')\n",
    "\n",
    "my_range=list(range(0,len(compiled_icb_rsid_plot[\"importance\"])))\n",
    "fig, ax = plt.subplots(figsize=(8,8))\n",
    "plt.hlines(y=my_range, xmin=0, xmax=compiled_icb_rsid_plot[\"importance\"], alpha=0.4, linewidth=7)\n",
    "plt.plot(compiled_icb_rsid_plot[\"importance\"], my_range, \"o\", markersize=7, alpha=0.7)\n",
    "# set labels style\n",
    "ax.set_xlabel('Relative Importance', fontsize=15, fontweight='black', color = '#333F4B')\n",
    "ax.set_ylabel('')\n",
    "# change the style of the axis spines\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "# ax.spines['left'].set_bounds((0, len(my_range)-1))\n",
    "ax.set_xlim(0,0.06)\n",
    "plt.yticks(range(len(compiled_icb_rsid_plot[\"importance\"])), compiled_icb_rsid_plot[\"rsid\"])\n",
    "# add some space between the axis and the plot\n",
    "ax.spines['left'].set_position(('outward', 8))\n",
    "ax.spines['bottom'].set_position(('outward', -10))\n",
    "\n",
    "plt.savefig(\"../plots/fig3/prostate_feature_importance_V2.pdf\",bbox_inches = \"tight\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "d241471e-ba7a-498f-822b-ed6ec79ad518",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  after removing the cwd from sys.path.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ellipse_outcome_cov_table_plot=ellipse_outcome_cov_table[ellipse_outcome_cov_table['LASSO_burden']<float(ellipse_outcome_cov_table['LASSO_burden'].quantile([0.99]))]\n",
    "minmax_scale = preprocessing.MinMaxScaler(feature_range=(0, 10))\n",
    "x = ellipse_outcome_cov_table_plot[\"LASSO_burden\"].values.reshape(-1, 1)\n",
    "ellipse_outcome_cov_table_plot[\"LASSO_scaled\"] = minmax_scale.fit_transform(x)\n",
    "\n",
    "fig = plt.figure(figsize=(5,5))\n",
    "\n",
    "#sns.boxplot(x=\"pheno\",y=\"LASSO_scaled\",data=ellipse_outcome_cov_table_plot)\n",
    "sns.violinplot(x=\"pheno\",y=\"LASSO_scaled\",data=ellipse_outcome_cov_table_plot)\n",
    "\n",
    "\n",
    "cat1 = ellipse_outcome_cov_table[ellipse_outcome_cov_table[\"pheno\"]==0]\n",
    "cat2 = ellipse_outcome_cov_table[ellipse_outcome_cov_table[\"pheno\"]==1]\n",
    "disc_score=cat2[\"LASSO_burden\"].mean()-cat1[\"LASSO_burden\"].mean()\n",
    "\n",
    "plt.title(stats.mannwhitneyu(cat1[\"LASSO_burden\"], cat2[\"LASSO_burden\"])[1])\n",
    "\n",
    "plt.ylabel(\"burden (pvalue < 0.05) RF Regression\")\n",
    "plt.xticks(rotation=90)\n",
    "plt.savefig(\"../plots/fig3/prostateViolin_0.99_quantile_Dropped_rescaled.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "e85e7925-5c97-457f-82d2-6ec0b082658e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#restructure PRS score so that it makes more sense to plot\n",
    "ellipse_outcome_cov_table[\"PRS\"]=ellipse_outcome_cov_table[\"LASSO_burden\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "b336d839-a0e6-4b3c-814f-ad447ad6a577",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9164\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.633312\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9163\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.633945\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9157\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.633066\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9158\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.630499\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9158\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.632314\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9155\n",
      "Name: group, dtype: int64\n",
      "1    9162\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.630556\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9156\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.626826\n",
      "         Iterations 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "/cellar/users/tsears/miniconda3/envs/ML/lib/python3.7/site-packages/ipykernel_launcher.py:27: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    9162\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.627650\n",
      "         Iterations 6\n",
      "1    9163\n",
      "0    9155\n",
      "Name: group, dtype: int64\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.628002\n",
      "         Iterations 6\n"
     ]
    }
   ],
   "source": [
    "ellipse_outcome_odds=quantile_plot(ellipse_outcome_cov_table,\"pheno ~ group + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10\",10,\"logistic\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "6cc939df-957c-4ba4-a039-109f7ab82b97",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f=plt.figure(figsize=(7,5))\n",
    "plt.errorbar(x=ellipse_outcome_odds[\"quant\"].tolist(),y=ellipse_outcome_odds[\"odds\"].tolist(),yerr=ellipse_outcome_odds[\"se\"].tolist(),fmt=\"o\",color=\"blue\",label=\"group1 + mhc genes\")\n",
    "plt.ylabel(\"quantile\")\n",
    "plt.ylabel(\"odds of C60-C63 \\n (prostate cancer)\")\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.savefig(\"../plots/fig3/ellipse.quantile.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "cc7c95f0-3346-4833-a0d6-94a1237bf81a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#1st vs 10th quantile\n",
    "ellipse_outcome_odds_filt=ellipse_outcome_odds[ellipse_outcome_odds[\"quant\"].isin([\"0.0\",\"90.0\"])]\n",
    "f=plt.figure(figsize=(3,5))\n",
    "plt.errorbar(x=ellipse_outcome_odds_filt[\"quant\"].tolist(),y=ellipse_outcome_odds_filt[\"odds\"].tolist(),yerr=ellipse_outcome_odds_filt[\"se\"].tolist(),fmt=\"o\")\n",
    "plt.xlabel(\"LASSO signature PRS quantile\")\n",
    "plt.ylabel(\"odds of prostate cancer\")\n",
    "plt.xticks([0,1])\n",
    "plt.xlim((-1, 2))\n",
    "plt.title(\"ELLIPSE Consortium\")\n",
    "plt.axhline(1,linestyle=\"--\",color=\"gray\")\n",
    "plt.savefig(\"../plots/fig3/1vs9_prostate.quantile.pdf\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "germtime",
   "language": "python",
   "name": "germtime"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}