Add files via upload

docs/scripts/tax_predict_vConTACT2_0.9.19.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python
+
+'''tax_predict_vConTACT2.0.11.3.py
+
+Generate viral taxonomy predictions for each contig based on guilt-by-association-clustering results from vConTACT2.
+
+- Developed based on output from vConTACT2.0.11.3.
+
+
+Optional parameters: 
+
+--vcontact2_results (-i) genome_by_genome_overview.csv : output file from vcontact2. Default = './genome_by_genome_overview.csv'
+--output_path (-o) output directory : Path for output directory. Default = './'
+'''
+
+from argparse import ArgumentParser
+import pandas as pd
+import numpy as np
+
+parser = ArgumentParser()
+parser.add_argument("-i", "--vcontact2_results", dest="infile", 
+                    help="Output file from vcontact2: genome_by_genome_overview.csv",
+                    metavar='vcontact2 genome_by_genome_overview.csv', required=True)
+parser.add_argument("-o", "--output_path", dest="outpath", 
+                    help="Path/to/output/directory. Default = current directory",
+                    metavar='output_path', default='.')
+args = parser.parse_args()
+
+def main():
+    print("\n--------------------\r\n")
+    print("Running tax_predict.py\r\n")
+    ## Read in vcontact2 results
+    df = pd.read_csv(args.infile)
+    df.columns = df.columns.str.replace(' ', '_')
+    # Add subcluster column, and strip subcluster string ('_n') from VC column
+    df['VC_Subcluster'] = df['VC']
+    df['VC'] = df['VC'].str.replace(r'(VC_.*)_.*', r'\1')
+    ## Generate taxonomy predictions for each viral cluster (one set for cluster, one set for subcluster)
+    # For each taxonomic rank: Groupby viral cluster (VC), generate list of unique taxonomies associated with that cluster (excluding 'Unassigned'), add <rank>_VC_predicted to each contig in that cluster.
+    # Replace instances of no predicted taxonomy for viral cluster (empty lists in predicted taxonomy column) with 'Unassigned'
+    for taxa in ['Order', 'Family', 'Genus']:
+        for VC_group in ['VC', 'VC_Subcluster']:
+            df = df.join(df.groupby(by=VC_group)[taxa].apply(lambda s: list({x for x in s if x != "Unassigned"})), on=VC_group, rsuffix='_'+VC_group+'_predicted')
+            df[taxa+'_'+VC_group+'_predicted'] = df[taxa+'_'+VC_group+'_predicted'].where(df[taxa+'_'+VC_group+'_predicted'].str.len() > 0, 'Unassigned')
+    ## Add taking into account p-values or confidence scores?
+    #
+    ## Write out tables
+    # genome_by_genome
+    df.to_csv(args.outpath+'/genome_by_genome_overview_tax_predictions.tsv', sep='\t', index=False)
+    # tax predictions
+    subset_columns = ['Genome', 'Order_VC_predicted', 'Family_VC_predicted', 'Genus_VC_predicted', 'Order_VC_Subcluster_predicted', 'Family_VC_Subcluster_predicted', 'Genus_VC_Subcluster_predicted', 'VC', 'VC_Subcluster', 'VC_Status']
+    df[subset_columns].to_csv(args.outpath+'/tax_predict_table.tsv', sep='\t', index=False)
+    print("Output:\r\n")
+    print(args.outpath+"/tax_predict_table.tsv:\nTable of predicted taxonomy for all contigs\r\n")
+    print(args.outpath+"/genome_by_genome_overview_tax_predictions.tsv:\nReproduced genome_by_genome_overview table with the addition of predicted taxonomy for all contigs\r\n")
+    print("Completed tax_predict.py\r\n")
+    print("\n--------------------\r\n")
+
+
+if __name__ == '__main__':
+    main()
+