def blastContextGenes(self,njobs=2):
print "Blasting Context Genes"
""" First split up the main bacteriocin file into a bunch of smaller files"""
split_fastafiles = ["%s/context.%d"%(self.intermediate,i)
for i in xrange(njobs)]
#self.split_files += split_fastafiles
split_fastahandles = [open(f,'w') for f in split_fastafiles]
out_classes = ["%s/contextout.%d"%(self.intermediate,i) for i in xrange(njobs)]
index=0
for record in SeqIO.parse(self.cand_context_genes_fasta,"fasta"):
if len(record.seq)<=1: continue #To weed out weird entries
split_fastahandles[index].write(">%s\n%s\n"%(str(record.id),
fasta.format(str(record.seq))))
index=(index+1)%njobs
#Close files
for handle in split_fastahandles: handle.close()
context_cmd = ' '.join([
"""module load anaconda; module load blast;module load blast+;""",
"""python %s/src/genome/context_gene.py""",
"""--training-directory=%s""",
"""--training-labels=%s""",
"""--query=%s""",
"""--intermediate=%s""",
"""--num-threads=%d""",
"""--output=%s"""
])
""" Release jobs """
jobs = []
for i in xrange(njobs):
cmd = context_cmd%(self.rootdir,
self.training_directory,
self.training_labels,
split_fastafiles[i],
self.intermediate,
self.numThreads,
out_classes[i]
)
batch_file = "%s/context_blast%i.%d.job"%(os.getcwd(),i,os.getpid())
self.batch_files.append(batch_file)
proc = quorum.Popen( cmd,shell=True,batch_file=batch_file,
stdin=quorum.PIPE,stdout=quorum.PIPE ,threads=self.numThreads)
proc.submit()
#proc.output = out_classes[i]
jobs.append(proc)
self.jobs.append(proc)
for job in jobs: job.wait()
""" Collect all of the results from the jobs"""
context_out = open(self.blast_context_out,'w')
for i in xrange(njobs):
if os.path.exists(out_classes[i]):
shutil.copyfileobj(open(out_classes[i]),context_out)
context_out.close()
pass
def reformat(orgfaa,faafile):
outhandle = open(faafile,'w')
for record in SeqIO.parse(open(orgfaa,'r'),'fasta'):
gi,num,dbtype,protid,_ = record.id.split("|")
outhandle.write(">%s\n%s\n"%(protid,fasta.format(str(record.seq))))
outhandle.close()
def getFasta(txt,fastadb,fastaindex,fastaout):
outhandle = open(fastaout,'w')
indexer = fasta.Indexer(fastadb,fastaindex)
indexer.load()
i = 0
with open(txt,'r') as handle:
for ln in handle:
ln = ln.rstrip()
toks = ln.split('|')
acc,clrname,full_evalue,hmm_st,hmm_end,env_st,env_end,description=toks
full_evalue = float(full_evalue)
hmm_st,hmm_end,env_st,env_end = map(int,[hmm_st,hmm_end,env_st,env_end])
seq = indexer.fetch(acc,env_st,env_end)
seq = fasta.format(seq)
outhandle.write(">%s:%d %s\n%s\n"%(acc,i,description,seq))
i+=1
def writeClusters(self,clusters,seq_dict,outhandle):
clusternum=0
for cluster in clusters:
for gene in cluster:
acc,clrname,full_evalue,hmm_st,hmm_end,env_st,env_end,description,strand,protid=gene.split("|")
hmm_st,hmm_end,env_st,env_end = map(int,[hmm_st,hmm_end,env_st,env_end])
full_evalue = float(full_evalue)
function = clrname.split('.')[0]
seq = seq_dict[(acc,clrname,full_evalue,hmm_st,hmm_end,env_st,env_end,description,strand,protid)]
outhandle.write(">accession=%s|function=%s|start=%s|end=%s|strand=%s|score=%s|protein_id=%s|cluster_%d|%s\n%s\n"%
(acc,function,env_st,env_end,strand,str(full_evalue),protid,clusternum,description,fasta.format(seq)))
clusternum+=1
