def overallCaptureFraction(dirinput, fileList, probes, region='intron'):
print('Reading overall captured base changes...')
from parseline import VCFObj
from collections import defaultdict
from Bio.Seq import Seq
vafs = {'C>A': [], 'C>G': [], 'C>T': [], 'T>A': [], 'T>C': [], 'T>G': []}
for i in fileList:
temp = {'C>A': 0, 'C>G': 0, 'C>T': 0, 'T>A': 0, 'T>C': 0, 'T>G': 0}
target = open(dirinput + '/' + i, 'r')
for line in target:
if '#' not in line and 'chr' in line: # skip the info
vcfObj = VCFObj(line)
# only use single substitutions and eliminate SNPs
if len(vcfObj.wt) == 1 and len(
vcfObj.var) == 1 and vcfObj.af < 0.1:
varType = ('%s>%s' % (vcfObj.wt, vcfObj.var))
if varType not in vafs:
varType = ('%s>%s' %
(str(Seq(vcfObj.wt).complement()),
str(Seq(vcfObj.var).complement())))
for i in probes:
for j in probes[i]:
if j in str(vcfObj.location):
if region == 'intron' and i[0] == 'T':
# add AO together
temp[varType] += int(vcfObj.ao)
elif region == 'exon' and not i[0] == 'T':
# add AO together
temp[varType] += int(vcfObj.ao)
target.close()
# calculate fractions
totalAO = 0
for i in temp:
totalAO += temp[i]
for i in temp:
temp[i] = float(temp[i]) / float(totalAO)
# transfer data into vafs
for i in temp:
vafs[i].append(temp[i])
return vafs
def readVCF(inFiles, inDir, ref, probes):
from parseline import VCFObj
from Bio.Seq import Seq
from getSequence import getRefSequence
from collections import defaultdict
totalAO = defaultdict(list)
print 'TotExon','ExonCount','TotIntr','IntrCount','Exonic','Intronic','NormExon','NormIntron'
for sample in inFiles:
totalExon = 0
totalIntron = 0
exonCount = 0
intronCount = 0
target = open(inDir + '/' + sample + '/onlyProbedRegions.vcf', 'r')
for line in target:
if '#' not in line and 'chr' in line: # skip the info
lobj = VCFObj(line)
# only C or G could be in CpG context
if (lobj.wt == 'C' or lobj.wt == 'G') and len(lobj.var) == 1 and lobj.af < 0.1:
seq = getRefSequence(line, 1, ref)
if seq[1:3] == 'CG' or seq[1:3] == 'GC':
for i in probes:
for loc in probes[i]:
if str(loc) in str(lobj.location) and i[0] == 'T':
exonCount += 1
totalExon += lobj.ao
elif str(loc) in str(lobj.location) and not i[0] == 'T':
intronCount += 1
totalIntron += lobj.ao
# first normalize by total number of sites in each category
exonic = totalExon / exonCount
intronic = totalIntron / intronCount
# normalize to 100 percent
normexonic = exonic / (exonic + intronic)
normintronic = intronic / (exonic + intronic)
print totalExon, exonCount, totalIntron, intronCount, exonic, intronic, normexonic, normintronic
# append normalized results
totalAO['Exonic'].append(normexonic)
totalAO['Intronic'].append(normintronic)
target.close()
print totalAO
return totalAO
def parseAll(inFiles, inDir, ref, probes):
from parseline import VCFObj
from Bio.Seq import Seq
from getSequence import getRefSequence
from collections import defaultdict
# allbases = {'C>A':{'intron':[0.4,0.5], 'exon':[0.6,0.5]}}
indiv = defaultdict(list)
allbases = {'C>A':indiv.copy(),'C>G':indiv.copy(),'C>T':indiv.copy(),'T>A':indiv.copy(),'T>C':indiv.copy(),'T>G':indiv.copy()}
for sample in inFiles:
exon = {'C>A':[0,0],'C>G':[0,0],'C>T':[0,0],'T>A':[0,0],'T>C':[0,0],'T>G':[0,0]}
intron = {'C>A':[0,0],'C>G':[0,0],'C>T':[0,0],'T>A':[0,0],'T>C':[0,0],'T>G':[0,0]}
target = open(inDir + '/' + sample + '/onlyProbedRegions.vcf', 'r')
for line in target:
if '#' not in line and 'chr' in line: # skip the info
lobj = VCFObj(line)
# only look at substitutions and non SNPs
if len(lobj.wt) == 1 and len(lobj.var) == 1 and lobj.af < 0.1:
label = ('%s>%s' % (lobj.wt, lobj.var))
if label not in exon:
wt = str(Seq(lobj.wt).complement())
var = str(Seq(lobj.var).complement())
label = ('%s>%s' % (wt, var))
for i in probes:
for loc in probes[i]:
if str(loc) in str(lobj.location) and i[0] == 'T':
intron[label][0] += lobj.ao
intron[label][1] += 1
elif str(loc) in str(lobj.location) and not i[0] == 'T':
exon[label][0] += lobj.ao
exon[label][1] += 1
# first normalize by total number of sites in each category
for i in exon:
exon[label][0] = exon[label][0] / exon[label][1]
intron[label][0] = intron[label][0] / intron[label][1]
# normalize to 100 percent
for i in exon:
normexonic = exon[i][0] / (exon[i][0] + intron[i][0])
normintronic = intron[i][0] / (exon[i][0] + intron[i][0])
allbases[i]['exon'].append(normexonic)
allbases[i]['intron'].append(normintronic)
target.close()
return allbases
def populatePandasDataframe(dirinput, fileList, probes, ref, upstream=10, downstream=10):
import pandas as pd
from Bio.Seq import Seq
print('Building data structure...')
allSamples = []
columns = ['Loc','WT','Var','Change','ConvChange','AO','DP','VAF','IntEx','Upstream','Downstream','Individual']
dat = []
tempAllVariants = []
sampleCount = 0
for sample in fileList:
inFile = open(dirinput + '/' + sample + '/onlyProbedRegions.vcf', 'r')
sampleCount += 1
for line in inFile:
if '#' not in line and 'chr' in line: # skip the info
lineobj = VCFObj(line)
# convert to six changes
if lineobj.wt == 'G' or lineobj.wt == 'A':
wt = str(Seq(lineobj.wt).complement())
var = str(Seq(lineobj.var).complement())
else:
wt = str(lineobj.wt)
var = str(lineobj.var)
surrounding = getRefSequence(lineobj, upstream, downstream, ref)
up = str(surrounding[:upstream])
down = str(surrounding[-downstream:])
probeRegion = ''
for probe in probes:
if len(probeRegion) < 1:
for loc in probes[probe]:
if str(loc) in str(lineobj.location):
if probe[0] == 'T':
probeRegion = 'TIII'
else:
probeRegion = 'Exon'
if len(lineobj.wt) == 1 and len(lineobj.var) == 1 and lineobj.af < 0.1:
dat = [lineobj.location, str(lineobj.wt), str(lineobj.var), str(lineobj.wt) + '>' + str(lineobj.var), wt + '>' + var, lineobj.ao, lineobj.dp, lineobj.af, probeRegion, up, down, sampleCount]
tempdat = pd.DataFrame(dat, index=columns)
tempAllVariants.append(tempdat.T)
inFile.close()
allVariants = pd.concat(tempAllVariants, ignore_index=True)
return allVariants
def parseAllVAF(inFiles, inDir, ref, probes):
from parseline import VCFObj
from Bio.Seq import Seq
from getSequence import getRefSequence
from collections import defaultdict
import numpy as np
# intron = {('CCG','C>A'):[0.75, 0.87]}
# exon = {('CCG','C>A'):[0.75, 0.87]}
intron = defaultdict(list)
exon = defaultdict(list)
for sample in inFiles:
tempintron = defaultdict(list)
tempexon = defaultdict(list)
target = open(inDir + '/' + sample + '/onlyProbedRegions.vcf', 'r')
for line in target:
if '#' not in line and 'chr' in line: # skip the info
lobj = VCFObj(line)
# only look at substitutions and non SNPs
if len(lobj.wt) == 1 and len(lobj.var) == 1 and lobj.af < 0.1:
if lobj.wt == 'C' or lobj.wt == 'T':
label = ('%s>%s' % (lobj.wt, lobj.var))
seq = getRefSequence(line, 1, ref)
else:
wt = str(Seq(lobj.wt).complement())
var = str(Seq(lobj.var).complement())
label = ('%s>%s' % (wt, var))
seq = str(Seq(getRefSequence(line, 1, ref)).reverse_complement())
for i in probes:
for loc in probes[i]:
if str(loc) in str(lobj.location) and i[0] == 'T':
tempintron[(seq,label)].append(lobj.af)
elif str(loc) in str(lobj.location) and not i[0] == 'T':
tempexon[(seq,label)].append(lobj.af)
for i in tempintron:
vafmean = np.mean(tempintron[i])
intron[i].append(vafmean)
for i in tempexon:
vafmean = np.mean(tempexon[i])
exon[i].append(vafmean)
target.close()
return intron, exon
def getSequence(vcfLine, flankLength):
from parseline import VCFObj
from subprocess import check_output, STDOUT
vcfObj = VCFObj(vcfLine)
low = int(vcfObj.location) - flankLength
high = int(vcfObj.location) + flankLength
temp = check_output(
'wget -qO- http://genome.ucsc.edu/cgi-bin/das/hg19/dna?segment=%s:%s,%s'
% (vcfObj.chrom, low, high),
stderr=STDOUT,
shell=True)
finalSeq = ''
for line in temp.split('\n'):
if '<' not in line:
finalSeq += line
return finalSeq
def getRefSequence(vcfLine, flankLength, ref):
from parseline import VCFObj
from subprocess import check_output, STDOUT
from string import upper
vcfObj = VCFObj(vcfLine)
low = int(vcfObj.location) - flankLength
high = int(vcfObj.location) + flankLength
temp = check_output('samtools faidx %s %s:%s-%s' %
(ref, vcfObj.chrom, low, high),
stderr=STDOUT,
shell=True)
finalSeq = ''
for line in temp.split('\n'):
if '>' not in line:
finalSeq += line
finalSeq = finalSeq.upper()
return finalSeq