def updateCons(path, db, motifChrom='chr17'):
"""update conservation scores in gzipped fixedStep wiggle format"""
mcollection = db["hg19"+motifChrom]
for infile in glob.glob(os.path.join(path, "*.wigFix.gz")):
(wigpath,wigfilename) = os.path.split(infile)
chrom = wigfilename.split('.')[0]
consName = '_'.join(wigfilename.split('.')[1:-2])
print 'updating', consName
#print chrom, tfName, consName
with gzip.open(infile) as wigFile:
#wig = csv.reader(wigFile,delimiter='\t')
bwFile = os.path.join(wigpath,consName+'.bw')
if not os.path.isfile(bwFile):
countWig.compressFixWig(wigFile, consName, bwFile)
stepDict, startDict, valuesDict = countWig.getBinFixStart(bwFile,consName)
start = startDict[consName][chrom]
arrayDict = countWig.buildFixHist(chrom,stepDict,startDict,valuesDict,consName)
cursor = mcollection.find()#{"tf_name": tfName,
#"genomic_region.chr": chrom})
print mcollection.count()
#num = 0
#avg = 0
for test in cursor:
motifStart, motifEnd = test["genomic_region"]["start"], test["genomic_region"]["end"]
#num += 1
#print avg#motifStart, motifEnd, num
avg = 0
startlist = [start[i] for i in xrange(len(start)-1) if (motifStart >= start[i] and motifStart < start[i+1]) or (motifEnd >= start[i] and motifEnd < start[i+1])]
if motifEnd > start[-1]:
startlist.append(start[-1])
for i in xrange(len(startlist)):
#if avg != 0:
#if motifEnd >= startlist[i]:##cases of partial overlap need to renormalize over two fragments
ss = startlist[i]
xs, xvals, sums, ll = arrayDict[ss]
if motifStart < ss <= motifEnd <= ss+ll-1:##left out, right in
if avg == 'NA' and i == len(startlist)-1:
avg = 0
avg += countWig.queryHist(xs,xvals, sums, ss, motifEnd)[0] *(motifEnd - ss + 1) /(motifEnd - motifStart + 1)
elif ss <= motifStart < motifEnd <= ss+ll-1:##in array
avg = countWig.queryHist(xs,xvals, sums, motifStart, motifEnd)[0]
elif motifStart < ss and ss+ll-1 < motifEnd:##motif > array
if avg == 'NA':
avg = 0
avg += countWig.queryHist(xs,xvals, sums, ss, ss+ll-1)[0] * ll /(motifEnd - motifStart + 1)
elif ss <= motifStart <= ss+ll-1 < motifEnd:##left in, right out
if avg == 'NA' and i == len(startlist)-1:
avg = 0
avg += countWig.queryHist(xs,xvals, sums, motifStart, ss+ll-1)[0] *(ss + ll - motifStart) /(motifEnd - motifStart + 1)
elif ss+ll-1 < motifStart:
if avg == 0:
#print 'here-->', motifStart, motifEnd, ss, ll
avg = 'NA'
elif motifEnd < ss:
print "this should not happen-- motifStart < motifEnd < ss "
if avg == 0:
avg = 'NA'
if avg != 'NA':
mcollection.update({"_id":test["_id"]},{"$set":{"cons."+consName: avg}}, upsert = True)
return 0
def getCons(path, tfName):
for infile in glob.glob(os.path.join(path, "*.wigFix")):
(wigpath,wigfilename) = os.path.split(infile)
chrom = wigfilename.split('.')[0]
consName = '_'.join(wigfilename.split('.')[1:3])
#print chrom, tfName, consName
with open(infile,'rt') as wigFile:
wig = csv.reader(wigFile,delimiter='\t')
stepDict, startDict, valuesDict = countWig.getFixStart(wig,consName)#'phyloP30wayEuarchontoglires')
start = startDict[consName][chrom]
arrayDict = countWig.buildFixHist(chrom,stepDict,startDict,valuesDict,consName)
cursor = mcollection.find({"tf_name": tfName,
"motif_genomic_regions_info.chr": chrom})
for test in cursor:
motifStart, motifEnd = test["motif_genomic_regions_info"]["start"], test["motif_genomic_regions_info"]["end"]
avg = 0
#print motifStart, motifEnd
startlist = [start[i] for i in xrange(len(start)-1) if (motifStart > start[i] and motifStart < start[i+1]) or (motifEnd > start[i] and motifEnd < start[i+1])]
if len(startlist) > 0:
print startlist
#print start[-1]
startlist.append(start[-1])
for i in xrange(len(startlist)):
#print arrayDict[start[i]]
# if avg != 0 and motifEnd < start[i]:
# break ##fall into range and break out
if avg != 0:
if motifEnd > startlist[i]:##cases of partial overlap need to renormalize over two fragments
xs, xvals, sums = arrayDict[startlist[i]]
avg = avg * (startlist[i] - motifStart) + (countWig.queryHist(xs,
xvals, sums, motifStart, motifEnd)[0] * (motifEnd - startlist[i] + 1)) / (motifEnd - motifStart + 1)
else:
break
elif avg == 0:
xs, xvals, sums = arrayDict[startlist[i]]
avg = countWig.queryHist(xs, xvals, sums, motifStart, motifEnd)[0]
if avg > 0:
print avg, motifStart, motifEnd
mcollection.update({"_id":test["_id"]},{"$set":{"motif_cons_info":{consName: avg}}}, upsert = True)
else:
mcollection.update({"_id":test["_id"]},{"$set":{"motif_cons_info":{consName: avg}}}, upsert = True)
#mcollection.save(test)
return 0
def updateCons(inpath, tfname, motifChrom, outpath):
"""update conservation scores in gzipped fixedStep wiggle format"""
#check if directories exists
motifdir = os.path.join(outpath,"bedMotifs")
consdir = os.path.join(outpath,"consMotifs")
if not os.path.isdir(motifdir):
print "Error: path-to-motif-bed-files invalid, please specify a valid outpath to store all calculated scores."
sys.exit()
#os.mkdir(motifdir)
if not os.path.isdir(consdir):
os.mkdir(consdir)
gcoordsfile = gzip.open(os.path.join(motifdir,tfname+motifChrom+".bed.gz"))
consfile = gzip.open(os.path.join(consdir,tfname+motifChrom+'cons.txt.gz'),'w')
writer = csv.writer(consfile, delimiter='\t')
l = []
consTypes = ["phastCons100way","phastCons46way","phastCons46way.placental","phastCons46way.primates",
"phyloP100way","phyloP46way","phyloP46way.placental","phyloP46way.primate"]
for consType in consTypes:
infile = os.path.join(inpath, motifChrom+"."+consType+".wigFix.gz")
(wigpath,wigfilename) = os.path.split(infile)
chrom = wigfilename.split('.')[0]
consName = '_'.join(wigfilename.split('.')[1:-2])
print 'updating', consName
gcoordsfile.seek(0)
gcoords = csv.reader(gcoordsfile, delimiter='\t')
with gzip.open(infile) as wigFile:
bwFile = os.path.join(wigpath,motifChrom+"."+consName+'.bw')
if not os.path.isfile(bwFile):
countWig.compressFixWig(wigFile, consName, bwFile)
stepDict, startDict, valuesDict = countWig.getBinFixStart(bwFile,consName)
start = startDict[consName][chrom]
arrayDict = countWig.buildFixHist(chrom,stepDict,startDict,valuesDict,consName)
r = []
for test in gcoords:
motifStart, motifEnd = int(test[1]), int(test[2])
#print motifStart, motifEnd
avg = 0
startlist = [start[i] for i in xrange(len(start)-1) if (motifStart >= start[i] and motifStart < start[i+1]) or (motifEnd >= start[i] and motifEnd < start[i+1])]
if motifEnd > start[-1]:
startlist.append(start[-1])
for i in xrange(len(startlist)):
#if avg != 0:
#if motifEnd >= startlist[i]:##cases of partial overlap need to renormalize over two fragments
ss = startlist[i]
xs, xvals, sums, ll = arrayDict[ss]
if motifStart < ss <= motifEnd <= ss+ll-1:##left out, right in
if avg == 'NA' and i == len(startlist)-1:
avg = 0
avg += countWig.queryHist(xs,xvals, sums, ss, motifEnd)[0] *(motifEnd - ss + 1) /(motifEnd - motifStart + 1)
elif ss <= motifStart < motifEnd <= ss+ll-1:##in array
avg = countWig.queryHist(xs,xvals, sums, motifStart, motifEnd)[0]
elif motifStart < ss and ss+ll-1 < motifEnd:##motif > array
if avg == 'NA':
avg = 0
avg += countWig.queryHist(xs,xvals, sums, ss, ss+ll-1)[0] * ll /(motifEnd - motifStart + 1)
elif ss <= motifStart <= ss+ll-1 < motifEnd:##left in, right out
if avg == 'NA' and i == len(startlist)-1:
avg = 0
avg += countWig.queryHist(xs,xvals, sums, motifStart, ss+ll-1)[0] *(ss + ll - motifStart) /(motifEnd - motifStart + 1)
elif ss+ll-1 < motifStart:
if avg == 0:
#print '...', motifStart, motifEnd, ss, ll
avg = 'NA'
elif motifEnd < ss:
print "Error: motifStart < motifEnd < ss "
if avg == 0:
avg = 'NA'
r.append(avg)
l.append(tuple(r))
wl = zip(*l)
for i in wl:
writer.writerows([list(i)])
consfile.close()
return 0
