Norm when abundance values are large

SemiBin/cluster.py

@@ -96,11 +96,12 @@ def run_embed_infomap(logger, model, data, * ,
     from scipy import sparse
     import torch
     import numpy as np
+    from .utils import norm_abundance
 
     train_data_input = data.values[:, 0:136] if not is_combined else data.values
 
     if is_combined:
-        if train_data_input.shape[1] - 136 > 20:
+        if norm_abundance(train_data_input):
             train_data_kmer = train_data_input[:, 0:136]
             train_data_depth = train_data_input[:, 136:len(data.values[0])]
             from sklearn.preprocessing import normalize

SemiBin/long_read_cluster.py

@@ -50,12 +50,13 @@ def cluster_long_read(logger, model, data, device, is_combined,
             n_sample, out, contig_dict, *, binned_length, args,
             minfasta):
     import pandas as pd
+    from .utils import norm_abundance
     contig_list = data.index.tolist()
     if not is_combined:
         train_data_input = data.values[:, 0:136]
     else:
         train_data_input = data.values
-        if train_data_input.shape[1] - 136 > 20:
+        if norm_abundance(train_data_input):
             train_data_kmer = train_data_input[:, 0:136]
             train_data_depth = train_data_input[:, 136:len(data.values[0])]
             from sklearn.preprocessing import normalize

SemiBin/self_supervised_model.py

@@ -4,6 +4,7 @@
 from torch.optim import lr_scheduler
 import sys
 from .semi_supervised_model import Semi_encoding_single, Semi_encoding_multiple, feature_Dataset
+from .utils import norm_abundance
 
 def loss_function(embedding1, embedding2, label):
     relu = torch.nn.ReLU()
@@ -70,7 +71,7 @@ def train_self(logger, out : str, datapaths, data_splits, is_combined=True,
             if not is_combined:
                 train_data = train_data[:, :136]
             else:
-                if train_data.shape[1] - 136 > 20:
+                if norm_abundance(train_data):
                     train_data_kmer  = train_data[:, :136]
                     train_data_depth = train_data[:, 136:]
                     train_data_depth = normalize(train_data_depth, axis=1, norm='l1')
@@ -82,7 +83,6 @@ def train_self(logger, out : str, datapaths, data_splits, is_combined=True,
                     train_data_split = np.concatenate((train_data_split_kmer, train_data_split_depth), axis = 1)
 
             data_length = len(train_data)
-
             # cannot link data is sampled randomly
             n_cannot_link = min(len(train_data_split) * 1000 // 2, 4_000_000)
             indices1 = np.random.choice(data_length, size=n_cannot_link)

SemiBin/semi_supervised_model.py

@@ -151,6 +151,7 @@ def train(logger, out, contig_fastas, binned_lengths, datas, data_splits, cannot
     import pandas as pd
     from sklearn.preprocessing import normalize
     import numpy as np
+    from .utils import norm_abundance
 
     train_data = pd.read_csv(datas[0], index_col=0).values
     if not is_combined:
@@ -215,7 +216,7 @@ def train(logger, out, contig_fastas, binned_lengths, datas, data_splits, cannot
                 train_data_input = train_data[:, 0:136]
                 train_data_split_input = train_data_must_link
             else:
-                if train_data.shape[1] - 136 > 20:
+                if norm_abundance(train_data):
                     train_data_kmer  = train_data[:, :136]
                     train_data_depth = train_data[:, 136:]
                     train_data_depth = normalize(train_data_depth, axis=1, norm='l1')

SemiBin/utils.py

@@ -631,3 +631,16 @@ def compute_min_length(min_length, fafile, ratio):
     if binned_short:
         return 1000
     return 2500
+
+def norm_abundance(data):
+    import numpy as np
+    n = data.shape[1] - 136
+    flag = False
+
+    if n >= 20:
+        flag = True
+    else:
+        if n >= 5:
+            if np.mean(np.sum(data[:, 136:], axis=1)) > 2:
+                flag = True
+    return flag