def parse_gpdfile(tdir,gpdfile,smoothing_factor):
# Go through the long reads and make a genepred
if gpdfile != '-':
fr = FileBasics.GenericFileReader(gpdfile)
else:
fr = sys.stdin
seennames = {}
longreadnumber = 0
of_gpd = open(tdir+'/longreads.gpd','w')
while True:
line = fr.readline()
if not line: break
if re.match('^#',line): #skip comments
continue
longreadnumber += 1
entry = GenePredBasics.smooth_gaps( \
GenePredBasics.line_to_entry(line.rstrip()) \
,smoothing_factor)
readname = entry['name']
if readname in seennames:
sys.stderr.write("Warning: repeat name '"+readname+"'\n")
#set our first name to our bin
entry['name'] = str(longreadnumber)
gline = GenePredBasics.entry_to_line(entry)
of_gpd.write(gline+"\n")
fr.close()
of_gpd.close()
def parse_refgpd(tdir,geneprednames,simplenames):
# get the reference genepreds ready to use in work
column_number = 0
entry_number = 0
of_entries = open(tdir+"/entries.txt",'w')
for file in geneprednames:
column_number += 1
of_ref = open(tdir+"/reference."+str(column_number)+".bed",'w')
gfr = FileBasics.GenericFileReader(file)
while True:
line = gfr.readline()
if not line: break
if re.match('^#',line): continue
entry_number += 1
line = line.rstrip("\n")
entry = GenePredBasics.line_to_entry(line)
entry_length = 0
for i in range(0,len(entry['exonStarts'])): entry_length += entry['exonEnds'][i]-entry['exonStarts'][i]
of_entries.write(str(column_number)+ "\t" + simplenames[column_number-1] + "\t" + str(entry_number) + "\t" + entry['gene_name'] + "\t" + entry['name']+"\t"+str(entry_length)+"\n")
exon_number = 0
for i in range(0,len(entry['exonStarts'])):
exon_number += 1
of_ref.write(entry['chrom'] + "\t" + str(entry['exonStarts'][i]) + "\t" \
+ str(entry['exonEnds'][i]) + "\t" + str(entry_number) + "\t" \
+ entry['gene_name'] + "\t" \
+ entry['name'] + "\t" + str(len(entry['exonStarts'])) + "\t" \
+ entry['strand'] + "\t" + str(exon_number) \
+ "\n")
gfr.close()
of_ref.close()
of_entries.close()
def gmap_all(reads_filename,index_base_name,outpsl_filename):
tdir = FileBasics.make_tempdir2('weirathe','gmap')
readformattag = ''
corecount = str(multiprocessing.cpu_count())
m = re.match('^(.*)\/([^\/]+)$',index_base_name)
if not m:
print "error: path should include both a directory and basename you should be able to use ./mybase if its in the same directory you are currently in"
sys.exit()
cmd = 'gmap -D '+m.group(1)+' -f 1 -d '+m.group(2)+' -t '+corecount+' '+reads_filename+' 1> '+tdir+'/all.psl 2>/dev/null'
sys.stderr.write(cmd+"\n")
os.system(cmd)
copyfile(tdir+'/all.psl',outpsl_filename)
rmtree(tdir)
def gmap_all(reads_filename, index_base_name, outpsl_filename):
tdir = FileBasics.make_tempdir2('weirathe', 'gmap')
readformattag = ''
corecount = str(multiprocessing.cpu_count())
m = re.match('^(.*)\/([^\/]+)$', index_base_name)
if not m:
print "error: path should include both a directory and basename you should be able to use ./mybase if its in the same directory you are currently in"
sys.exit()
cmd = 'gmap -D ' + m.group(1) + ' -f 1 -d ' + m.group(
2
) + ' -t ' + corecount + ' ' + reads_filename + ' 1> ' + tdir + '/all.psl 2>/dev/null'
sys.stderr.write(cmd + "\n")
os.system(cmd)
copyfile(tdir + '/all.psl', outpsl_filename)
rmtree(tdir)
def read_fastq_file(self,filename):
gfr = FileBasics.GenericFileReader(filename)
linecount = 0
while True and linecount < self.max_read_count:
line1 = gfr.readline().rstrip()
if not line1: break
line2 = gfr.readline().rstrip()
if not line2: break
line3 = gfr.readline().rstrip()
if not line3: break
line4 = gfr.readline().rstrip()
if not line4: break
self.record_observation(line4)
linecount += 1
gfr.close()
def main():
if len(sys.argv) != 5:
print sys.argv[0]+' <long reads fasta> <reference genome> <transcriptome genepred>'
return
longreadfname = sys.argv[1]
genomefname = sys.argv[2]
usergenepredfname = sys.argv[3]
outbase = sys.argv[4]
#get read count
readcount = 0
with open(longreadfname) as f:
for line in f:
if re.match('^>',line): readcount+=1
tdir = FileBasics.make_tempdir2('weirathe','annlong')
# 1. Make sure the transcriptome is uniquely named uniquely mapped entries
genepredfname = tdir+'/txn.gpd'
make_unique_genepred(usergenepredfname,genepredfname)
print 'made unique genepred file'
# 2. Make a transcriptome to align to.
transcriptomefasta = tdir+'/txn.fa'
genepred_basics.write_genepred_to_fasta_directionless(genepredfname,genomefname,transcriptomefasta)
print 'made transcriptome fasta'
# 3. Make a bed file of junction locations in that transcriptome.
junctionbedfname = tdir+'/junction.bed'
junction_counts = make_junction_bed_file(genepredfname,junctionbedfname)
print 'made junction bed file'
# 4. Build a gmap index of the transcriptome
transcriptomeindex = tdir+'/gmap_txn'
aligner_basics.build_gmap_index(transcriptomefasta,transcriptomeindex)
print 'made gmap index of transcriptome'
# 5. Align the long reads to the transcriptome with gmap
alignmentfname = tdir+'/reads.psl'
aligner_basics.gmap_all(longreadfname,transcriptomeindex,alignmentfname)
print 'made gmap alignment of reads to transcriptome'
# 6. Generate get the genepred of the long reads on the transcriptome coordinates.
# Smooth that genepred by a smoothing factor
# And make a bed file of the best alignment. see function for specifications
bestalignmentbedfname = tdir+'/reads.bed'
make_best_continuous_alignment_bed(alignmentfname,bestalignmentbedfname)
print 'made best continuous alignment bed file'
# 10. Print per-gene count info
genenames = genepred_basics.get_transcript_to_gene_name_dictionary(genepredfname)
print 'got gene name conversions'
# 7. Make a report of all prefilter alignments
bestprefilter = tdir+'/prefilter.txt'
prefilter_alignments = make_best_alignment_summary(bestalignmentbedfname,junctionbedfname,junction_counts,bestprefilter)
print 'made best alignment prefilter summary'
report_file = tdir +'/report.txt'
orep = open(report_file,'w')
orep.write('Basename:'+"\t"+outbase+"\n")
orep.write('Temp directory:'+"\t"+tdir+"\n")
orep.write('Long Read Count:'+"\t"+str(readcount)+"\n")
# 8. Filter the full length alignments
full_length_alignments = filter_alignments(prefilter_alignments,'full')
full_length_alignment_file = tdir+'/full_length_alignment.txt'
[full_length_read_count, full_length_transcript_count] = write_alignments(full_length_alignments,full_length_alignment_file,genenames)
orep.write('Read count - full length reads mapped:'+"\t" +str(len(full_length_alignments))+"\n")
orep.write('Transcript count - full length reads mapped:'+"\t" +str(full_length_transcript_count)+"\n")
unambiguous_full_length_alignment_file = tdir+'/unambiguous_full_length_alignment.txt'
unambiguous_full_length_alignments = filter_unambiguous_alignments(full_length_alignments)
[unambiguous_full_length_read_count, unambiguous_full_length_transcript_count] = write_alignments(unambiguous_full_length_alignments,unambiguous_full_length_alignment_file,genenames)
orep.write('Read count - full length reads mapped with unambiguous matches:'+"\t"+str(len(unambiguous_full_length_alignments))+"\n")
orep.write('Transcript count - full length reads mapped with unambiguous matches:'+"\t"+str(unambiguous_full_length_transcript_count)+"\n")
# 9. Filter the full length alignments
prepartial_alignments = filter_alignments(prefilter_alignments,'partial')
prepartial_alignment_file = tdir+'/prepartial_alignment.txt'
write_alignments(prepartial_alignments,prepartial_alignment_file,genenames)
partial_alignments = filter_by_priority_alignments(prepartial_alignments)
partial_alignment_file = tdir+'/partial_alignment.txt'
[partial_read_count, partial_transcript_count] = write_alignments(partial_alignments,partial_alignment_file,genenames)
orep.write('Read count - reads mapped with partial hits best junction and length matches:'+"\t" + str(len(partial_alignments))+"\n")
orep.write('Transcript count - reads mapped with partial hits best junction and length matches:'+"\t" + str(partial_transcript_count)+"\n")
unambiguous_partial_alignments = filter_unambiguous_alignments(partial_alignments)
unambiguous_partial_alignment_file = tdir + '/unambiguous_partial_alignments.txt'
[unambiguous_partial_read_count, unambiguous_partial_transcript_count] = write_alignments(unambiguous_partial_alignments,unambiguous_partial_alignment_file,genenames)
orep.write('Read count - reads mapped with partial hits unambiguous matches:'+"\t"+str(len(unambiguous_partial_alignments))+"\n")
orep.write('Transcript count - reads mapped with partial hits unambiguous matches:'+"\t"+str(unambiguous_partial_transcript_count)+"\n")
partial_gene_counts = get_uniquely_mappable_gene_counts(partial_alignments,genenames)
partial_gene_counts_file = tdir+'/partial_match_uniquely_mappable_gene_counts.txt'
write_gene_counts(partial_gene_counts,partial_gene_counts_file)
full_gene_counts = get_uniquely_mappable_gene_counts(full_length_alignments,genenames)
full_gene_counts_file = tdir+'/full_length_match_uniquely_mappable_gene_counts.txt'
write_gene_counts(full_gene_counts,full_gene_counts_file)
orep.write('Gene count - full length matches uniquely mapped:'+"\t"+str(len(full_gene_counts))+"\n")
orep.write('Gene count - partial matches uniquely mapped:'+"\t"+str(len(partial_gene_counts))+"\n")
orep.close()
copyfile(report_file,outbase+'.Report.txt')
copyfile(full_gene_counts_file,outbase+'.FullGeneCounts.txt')
copyfile(partial_gene_counts_file,outbase+'.PartialGeneCounts.txt')
copyfile(full_length_alignment_file,outbase+'.FullAlignment.txt')
copyfile(unambiguous_full_length_alignment_file,outbase+'.UnambiguousFullAlignment.txt')
copyfile(partial_alignment_file,outbase+'.PartialAlignment.txt')
copyfile(unambiguous_partial_alignment_file,outbase+'.UnambiguousFullAlignment.txt')
rmtree(tdir)