kkonganti@92: #!/usr/bin/env python kkonganti@92: kkonganti@92: import os kkonganti@92: import argparse kkonganti@92: import logging as log kkonganti@92: import pandas as pd kkonganti@92: import numpy as np kkonganti@92: from Bio import SeqIO kkonganti@92: kkonganti@92: kkonganti@92: def main(): kkonganti@92: # READ IN ARGUMENTS kkonganti@92: desc = """ kkonganti@92: This script is part of the centriflaken pipeline: It processes centrifuge kkonganti@92: output and produces either a filtered FASTQ or a text file of FASTQ IDs based kkonganti@92: on the supplied taxa/bug kkonganti@92: """ kkonganti@92: parser = argparse.ArgumentParser(prog='process_centrifuge_output.py', description=desc) kkonganti@92: parser.add_argument("-v", dest='verbose', action="store_true", help="For more verbose output") kkonganti@92: parser.add_argument("-i", dest='input_fastq', required=False, kkonganti@92: help="Path to input FASTQ file (same as input to centrifuge). If not mentioned, \ kkonganti@92: a text file of sequence IDs are produced instead of a FASTQ file") kkonganti@92: parser.add_argument("-t", dest='taxa_filtered_fastq_file', required=True, kkonganti@92: help="Path to output FASTQ or output text file filtered by the taxa specified") kkonganti@92: parser.add_argument("-r", dest='cent_report', required=True, help="Path to centrifuge report") kkonganti@92: parser.add_argument("-o", dest='cent_output', required=True, help="Path to centrifuge output") kkonganti@92: parser.add_argument("-b", dest='bug', required=True, kkonganti@92: help="Name or fragment of name of the bug by which reads are extracted") kkonganti@92: args = parser.parse_args() kkonganti@92: kkonganti@92: # MORE INFO IF VERBOSE kkonganti@92: if args.verbose: kkonganti@92: log.basicConfig(format="%(levelname)s: %(message)s", level=log.DEBUG) kkonganti@92: else: kkonganti@92: log.basicConfig(format="%(levelname)s: %(message)s") kkonganti@92: kkonganti@92: # ASSIGN VARIABLES kkonganti@92: input_fastq = args.input_fastq kkonganti@92: taxa_filtered_fastq_file = args.taxa_filtered_fastq_file kkonganti@92: cent_report = args.cent_report kkonganti@92: cent_output = args.cent_output kkonganti@92: bug = args.bug kkonganti@92: report_col_list = ["name", "taxID"] kkonganti@92: output_col_list = ["taxID", "readID"] kkonganti@92: kkonganti@92: # Match and filter taxa names and ids from centrifuge report file kkonganti@92: report_df = pd.read_csv(cent_report, delimiter="\t", usecols=report_col_list) kkonganti@92: report_df['name'] = report_df['name'].str.lower() kkonganti@92: filt_report_df = report_df[report_df['name'].str.contains(bug.lower())] kkonganti@92: #print("\nMatching taxa names and ids:\n",filt_report_df) kkonganti@92: taxID_list = filt_report_df['taxID'] kkonganti@92: kkonganti@92: # Match the above tax ids to read ids from centrifuge output file and deduplicate kkonganti@92: output_df = pd.read_csv(cent_output, delimiter="\t", usecols=output_col_list) kkonganti@92: filt_output_df = output_df.loc[output_df['taxID'].isin(taxID_list)] kkonganti@92: readID_list = filt_output_df['readID'] kkonganti@92: readID_dedup_list = np.unique(readID_list) kkonganti@92: TF=open(taxa_filtered_fastq_file, "w") kkonganti@92: kkonganti@92: if (not input_fastq): kkonganti@92: # print("\nFILTERED READ ID LIST:\n", readID_dedup_list) kkonganti@92: for ID in readID_dedup_list: kkonganti@92: TF.write(f"{ID}\n") kkonganti@92: else: kkonganti@92: # Extract filtered reads from input fastq and write to output fastq kkonganti@92: print ("Indexing reads..") kkonganti@92: rec = SeqIO.index(input_fastq,"fastq") kkonganti@92: for i in readID_dedup_list: kkonganti@92: if i in rec: kkonganti@92: SeqIO.write(rec[i], TF, "fastq") kkonganti@92: kkonganti@92: TF.close() kkonganti@92: kkonganti@92: if __name__ == "__main__": kkonganti@92: main()