annotate sopra_wpc.py @ 0:988d5a82291a draft

Uploaded

# -*- coding: utf-8 -*-
"""
SOPRA with prebuilt contigs workflow runner
"""

import optparse
import os
import tempfile
import shutil
import subprocess
import sys


# Copyright (c) Twisted Matrix Laboratories.
def which(name, flags=os.X_OK):
    """ Search PATH for executable files with the given name. """
    result = []
    exts = filter(None, os.environ.get('PATHEXT', '').split(os.pathsep))
    path = os.environ.get('PATH', None)
    if path is None:
        return []
    for p in os.environ.get('PATH', '').split(os.pathsep):
        p = os.path.join(p, name)
        if os.access(p, flags):
            result.append(p)
        for e in exts:
            pext = p + e
            if os.access(pext, flags):
                result.append(pext)
    return result


def __main__():
    parser = optparse.OptionParser(description='SOPRA with prebuilt contigs')
    parser.add_option('--contigs', action='append', dest='contigs', help='Contigs FASTA files, at least 1')
    parser.add_option('--mate', action='append', dest='mates', help='Paired-end Illumina libraries, at least 1 FASTA file')
    parser.add_option('-d', action='append', dest='insert_sizes', type='int', help='List of insert sizes for the corresponding mate pair libraries')
    parser.add_option('-v', dest='max_mismatches', type='int', help='Maximum number of mismatches when aligning reads on contigs with Bowtie')
    parser.add_option('-c', dest='c_option', type='int', help='If the number of times a read and its reverse complement appear in the library is equal to or more than this value, the pairing information from that read will be disregarded')
    parser.add_option('-w', dest='w_option', type='int', help='Minimum number of links between two contigs')
    parser.add_option('-L', dest='L_option', type='int', help='Minimum length of contigs to be used in scaffold assembly')
    parser.add_option('--h_option', dest='h_option', type='float', help='High coverage contigs (above mean coverage + h x std coverage) are not considered in the scaffold assembly mainly to exclude reads from repetitive regions')
    parser.add_option('--scaffolds', dest='scaffolds', help='scaffolds fasta file mandatory')
    parser.add_option('-l', '--logfile', dest='logfile', help='log file (default=stdout)')
    (options, args) = parser.parse_args()
    if len(args) > 0:
        parser.error('Wrong number of arguments')

    contigs = options.contigs # a list of file paths
    mates = options.mates # a list of file paths
    insert_sizes = options.insert_sizes # a list of integers
    max_mismatches = options.max_mismatches
    c_option = options.c_option
    w_option = options.w_option
    L_option = options.L_option
    h_option = options.h_option
    scaffolds = options.scaffolds
    logfile = options.logfile

    s_scaf_path = which('s_scaf_v1.4.6.pl').pop()
    print 'Creating temp dir'
    wd = tempfile.mkdtemp()
    try:
        fake_mates = [os.path.join(wd, os.path.basename(mate) + '.fasta') for mate in mates] # s_prep_contigAseq_v1.4.6.pl wants a mate file with extension [Ff][Aa][Ss][Tt][Aa] or [Ff][Aa]
        contigs_sopra = os.path.join(wd, 'contigs_sopra.fasta') # s_prep_contigAseq_v1.4.6.pl always writes all the prepared contigs to this file
        bowtie_build = os.path.join(wd, 'bowtie_build') # arbitrary basename for bowtie-build output files
        mate_sopras = [os.path.splitext(fake_mate)[0] + '_sopra.fasta' for fake_mate in fake_mates] # s_prep_contigAseq_v1.4.6.pl writes the prepared paired reads to these files
        mysam_mates = [mate_sopra + '.sam' for mate_sopra in mate_sopras] # arbitrary filenames for bowtie output in SAM format
        mysam_mates_parsed = [mysam_mate + '_parsed' for mysam_mate in mysam_mates] # s_parse_sam_v1.4.6.pl writes its output to these files
        orientdistinfo = os.path.join(wd, 'orientdistinfo_c%d' % c_option) # s_read_parsed_sam_v1.4.6.pl writes its output to this file
        scaffolds_file = os.path.join(wd, "scaffolds_h%s_L%d_w%d.fasta" % (h_option, L_option, w_option)) # s_scaf_v1.4.6.pl writes its output to this file

        for i in range(len(mates)):
            print "Copying mate %s to %s" % (mates[i], fake_mates[i])
            shutil.copy2(mates[i], fake_mates[i])

        log = open(logfile, 'w') if logfile else sys.stdout
        try:
            cmd_step1 = "s_prep_contigAseq_v1.4.6.pl -contig %s -mate %s -a %s" % (" ".join(contigs), " ".join(fake_mates), wd)
            print "SOPRA with prebuilt contigs (preparation) command to be executed:\n %s" % cmd_step1
            subprocess.check_call(args=cmd_step1, stdout=log, shell=True)

            cmd_step2 = "bowtie-build %s %s" % (contigs_sopra, bowtie_build)
            print "SOPRA with prebuilt contigs (Bowtie building index) command to be executed:\n %s" % cmd_step2
            subprocess.check_call(args=cmd_step2, stdout=log, shell=True)

            for i in range(len(mate_sopras)):
                cmd_step3 = "bowtie -v %d -m 1 -f --sam %s %s %s" % (max_mismatches, bowtie_build, mate_sopras[i], mysam_mates[i])
                print "SOPRA with prebuilt contigs (Bowtie alignment of library %d) command to be executed:\n %s" % (i+1, cmd_step3)
                subprocess.check_call(args=cmd_step3, stdout=log, stderr=subprocess.STDOUT, shell=True) # need to redirect stderr because bowtie writes some logging info there

            cmd_step4 = "s_parse_sam_v1.4.6.pl -sam %s -a %s" % (' '.join(mysam_mates), wd)
            print "SOPRA with prebuilt contigs (removing reads not mapped in a proper pair) command to be executed:\n %s" % cmd_step4
            subprocess.check_call(args=cmd_step4, stdout=log, shell=True)

            cmd_step5 = "s_read_parsed_sam_v1.4.6.pl -c %d -a %s" % (c_option, wd)
            for i in range(len(mysam_mates_parsed)):
                cmd_step5 += " -parsed %s -d %d" % (mysam_mates_parsed[i], insert_sizes[i])
            print "SOPRA with prebuilt contigs (read parsed SAM) command to be executed:\n %s" % cmd_step5
            subprocess.check_call(args=cmd_step5, stdout=log, shell=True)

            cmd_step6 = "perl -X %s -w %d -L %d -h %s -o %s -a %s" % (s_scaf_path, w_option, L_option, h_option, orientdistinfo, wd) # need to call with perl -X because: 1) otherwise some Perl warnings are written on stderr; 2) simply redirecting stderr would hide real errors since it always returns exit status 0
            print "SOPRA with prebuilt contigs (scaffold assembly) command to be executed:\n %s" % cmd_step6
            subprocess.check_call(args=cmd_step6, stdout=log, shell=True)
        finally:
            if log != sys.stdout:
                log.close()

        print 'Moving result file %s to %s' % (scaffolds_file, scaffolds)
        shutil.move(scaffolds_file, scaffolds)
    finally:
        shutil.rmtree(wd)


if __name__ == "__main__":
    __main__()