merge_bcs_reads.py

#!/usr/bin/env python

from sys import argv
from itertools import izip
 
from cogent.parse.fastq import MinimalFastqParser

""" Usage
python merge_bcs_reads.py X Y Z
X: barcodes fastq file
Y: reads fastq file
Z: merged output file

Will write the barcodes at the beginning of the output file
"""

 
bcs = open(argv[1], "U")
reads = open(argv[2], "U")
combined_out = open(argv[3], "w")

header_index = 0
sequence_index = 1
quality_index = 2 
 
for bc_data,read_data in izip(MinimalFastqParser(bcs,strict=False),
                               MinimalFastqParser(reads,strict=False)):
   
   curr_label = bc_data[header_index].strip()
   bc_read = bc_data[sequence_index]
   bc_qual = bc_data[quality_index]
   seq_read = read_data[sequence_index]
   seq_qual = read_data[quality_index]
   
   combined_out.write("@%s\n" % curr_label)
   combined_out.write("%s\n" % (bc_read + seq_read))
   combined_out.write("+\n")
   combined_out.write("%s\n" % (bc_qual + seq_qual))