def main_mergePep(args, stdout, stderr):
msg = "Building the hash index"
stderr.write(msg + "\n")
hashIndex = pygenes.buildGeneTableFileIndex(args.input)
msg = "Building the length index"
stderr.write(msg + "\n")
lengthIndex = pygenes.buildLengthFileIndex(hashIndex)
hashToMerged = dict()
with open(args.fasta, "w") as fo:
for l in lengthIndex.keys():
msg = "Merging sequences of length " + str(l)
stderr.write(msg + "\n")
sequences = pygenes.gatherSequences(args.input, lengthIndex, l)
mergedSequences = pygenes.mergeSequences(sequences, args.dissim)
for (k, v) in mergedSequences.items():
originalHash = pygenes.md5hash(k)
mergedHash = pygenes.md5hash(v)
assert not hashToMerged.get(originalHash, False)
hashToMerged[originalHash] = mergedHash
newMerged = set(mergedSequences.values())
for seq in newMerged:
fo.write(">" + pygenes.md5hash(seq) + "\n")
fo.write(seq + "\n")
with open(args.input, "r") as fi:
with open("tmp." + args.output, "w") as fo:
headers = fi.readline()
headerElements = headers.lstrip("#").strip().split("\t")
fo.write(headers)
for line in fi:
content = dict(zip(headerElements, line.strip().split("\t")))
content["mergedPeptideHash"] = hashToMerged[content["peptideHash"]]
fo.write("\t".join([content[x] for x in headerElements]) + "\n")
shutil.move("tmp." + args.output, args.output)
def main_splitHclust(args, stdout, stderr) :
# Load conservation file and determine files to realign
alnToSplit = set([])
with open(args.conservation, "r") as fi :
for l in fi :
if l.strip() != "" :
fasta, cons = l.strip().split("\t")
if float(cons) < args.threshold :
alnToSplit.add(fasta)
# Go through the input files
if args.outDir is None :
args.outDir = "."
processedFile = 0
for fastaFile in args.input :
total = str(len(alnToSplit))
if os.path.basename(fastaFile) in alnToSplit :
processedFile += 1
stderr.write("Processing file " + os.path.basename(fastaFile) +
" " + str(processedFile) + "/" + total + "\n")
# Build mapping from peptide sequences to sequence names
seqParser = SeqIO.parse(fastaFile, "fasta")
seqRaw = [x for x in seqParser]
seqs = dict()
[seqs.update({x.description : str(x.seq)}) for x in seqRaw]
pep2seqNames = collections.defaultdict(lambda : [])
[pep2seqNames[v].append(k) for (k, v) in seqs.iteritems()]
# Produce merged sequences
uniqueSeqs = list(set(seqs.values()))
stderr.write("Working with " + str(len(uniqueSeqs)) +
" unique sequences\n")
if (args.unique is None) or (len(uniqueSeqs) <= args.unique) :
mergedSeqs = pygenes.mergeSequences(uniqueSeqs,
maxDistance = args.dissim,
stderr = stderr)
# Build mapping from merged sequences to original peptide sequences
merged2pep = collections.defaultdict(lambda : [])
[merged2pep[v].append(k) for (k, v) in mergedSeqs.iteritems()]
# Output
for (i, v) in enumerate(merged2pep.values()) :
outFile = os.path.join(args.outDir,
(os.path.basename(fastaFile) + ".split" +
str(i) + ".fa"))
with open(outFile, "w") as fo :
for originalPep in v :
for seqName in pep2seqNames[originalPep] :
fo.write(">" + seqName + "\n")
fo.write(originalPep + "\n")
if args.outDir == "." and not args.keep :
os.remove(fastaFile)
else :
stderr.write("Too many unique sequences! File not processed\n")
if args.outDir == "." and not args.keep :
stderr.write("File deleted\n")
os.remove(fastaFile)
else :
if args.outDir != "." :
shutil.copy(fastaFile,
os.path.join(args.outDir,
os.path.basename(fastaFile)))