def extractSVReadPairs(bamFilePath, outputFilePath, juncChr1, juncPos1,
juncDir1, juncChr2, juncPos2, juncDir2, max_depth,
search_length, search_margin):
"""
read pairs containing break points are extracted. (yshira 2015/04/23)
The exact condition is as follows:
1. one of the read in the pair has the break point of the SV candidate
2. the start positions of the read pairs are within 800bp of the break point of the SV candidate
Some minor concern for the above conditions are:
1. Depending on the choice of the "start position" or "end position", the distance between the read and break point differs. This can generate slight bias...
(but I believe we can recover this by setting sufficient margin (800bp), and summarize the alignment result carefully.)
2. Maybe, for some of the read pair, the result of alignment is obvious. But should we re-align them?
"""
bamfile = pysam.Samfile(bamFilePath, 'rb')
# if the #sequence read is over the `maxDepth`, then that key is ignored
depthFlag = 0
if bamfile.count(juncChr1,
int(juncPos1) - 1,
int(juncPos1) + 1) >= max_depth:
depthFlag = 1
if bamfile.count(juncChr2,
int(juncPos2) - 1,
int(juncPos2) + 1) >= max_depth:
depthFlag = 1
if depthFlag == 1:
print >> sys.stderr, "sequence depth exceeds the threshould for: " + ','.join(
[juncChr1, juncPos1, juncDir1, juncChr2, juncPos2, juncDir2])
return 1
hOUT = open(outputFilePath, 'w')
readID2exist = {}
for read in bamfile.fetch(juncChr1, max(0,
int(juncPos1) - search_length),
int(juncPos1) + search_length):
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
chr_current = bamfile.getrname(read.tid)
pos_current = int(read.pos + 1)
dir_current = ("-" if flags[4] == "1" else "+")
chr_pair = bamfile.getrname(read.rnext)
pos_pair = int(read.pnext + 1)
dir_pair = ("-" if flags[5] == "1" else "+")
# the read (with margin) contains break point
if pos_current - search_margin <= int(juncPos1) <= (read.aend -
1) + search_margin:
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr1 and pos_current <= int(
juncPos1
) <= pos_pair and dir_current == "+" and dir_pair == "-":
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr2:
juncFlag = 0
if juncDir1 == "+" and juncDir2 == "+" and pos_current <= int(
juncPos1) and pos_pair <= int(juncPos2):
juncFlag = 1
if juncDir1 == "+" and juncDir2 == "-" and pos_current <= int(
juncPos1) and pos_pair >= int(juncPos2):
juncFlag = 1
if juncDir1 == "-" and juncDir2 == "+" and pos_current >= int(
juncPos1) and pos_pair <= int(juncPos2):
juncFlag = 1
if juncDir1 == "-" and juncDir2 == "-" and pos_current >= int(
juncPos1) and pos_pair >= int(juncPos2):
juncFlag = 1
if juncFlag == 1:
readID2exist[read.qname] = 1
for read in bamfile.fetch(juncChr2, max(0,
int(juncPos2) - search_length),
int(juncPos2) + search_length):
if read.qname == "ST-E00104:162:H03UUALXX:5:1222:21168:16006":
pass
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
chr_current = bamfile.getrname(read.tid)
pos_current = int(read.pos + 1)
dir_current = ("-" if flags[4] == "1" else "+")
chr_pair = bamfile.getrname(read.rnext)
pos_pair = int(read.pnext + 1)
dir_pair = ("-" if flags[5] == "1" else "+")
# the read (with margin) contains break point
if pos_current - search_margin <= int(juncPos2) <= (read.aend -
1) + search_margin:
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr2 and pos_current <= int(
juncPos2
) <= pos_pair and dir_current == "+" and dir_pair == "-":
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr1:
juncFlag = 0
if juncDir2 == "+" and juncDir1 == "+" and pos_current <= int(
juncPos2) and pos_pair <= int(juncPos1):
juncFlag = 1
if juncDir2 == "+" and juncDir1 == "-" and pos_current <= int(
juncPos2) and pos_pair >= int(juncPos1):
juncFlag = 1
if juncDir2 == "-" and juncDir1 == "+" and pos_current >= int(
juncPos2) and pos_pair <= int(juncPos1):
juncFlag = 1
if juncDir2 == "-" and juncDir1 == "-" and pos_current >= int(
juncPos2) and pos_pair >= int(juncPos1):
juncFlag = 1
if juncFlag == 1:
readID2exist[read.qname] = 1
readID2seq1 = {}
readID2seq2 = {}
complement = {'A': 'T', 'C': 'G', 'G': 'C', 'T': 'A', 'N': 'N'}
for read in bamfile.fetch(juncChr1, max(0,
int(juncPos1) - search_length),
int(juncPos1) + search_length):
if read.qname in readID2exist:
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
tempSeq = ""
if flags[4] == "1":
tempSeq = utils.reverseComplement(str(read.seq))
else:
tempSeq = read.seq
# the first read
if flags[6] == "1":
readID2seq1[read.qname] = tempSeq
else:
readID2seq2[read.qname] = tempSeq
for read in bamfile.fetch(juncChr2, max(0,
int(juncPos2) - search_length),
int(juncPos2) + search_length):
if read.qname in readID2exist:
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
tempSeq = ""
if flags[4] == "1":
tempSeq = utils.reverseComplement(str(read.seq))
else:
tempSeq = read.seq
# the first read
if flags[6] == "1":
readID2seq1[read.qname] = tempSeq
else:
readID2seq2[read.qname] = tempSeq
for readID in readID2seq1:
if readID in readID2seq2:
print >> hOUT, '>' + readID + '/1'
print >> hOUT, readID2seq1[readID]
print >> hOUT, '>' + readID + '/2'
print >> hOUT, readID2seq2[readID]
bamfile.close()
hOUT.close()
return 0
def clusterJunction(inputFilePath, outputFilePath, check_margin_size):
"""
script for merging and summarizing junction read pairs
"""
hIN = open(inputFilePath, 'r')
hOUT = open(outputFilePath, 'w')
mergedBedpeInfo = {}
mergedJunction = {}
for line in hIN:
F = line.rstrip('\n').split('\t')
match = 0
delList = []
for key in sorted(mergedBedpeInfo):
tchr1, tstart1, tend1, tchr2, tstart2, tend2, tdir1, tdir2, inseqSize = key.split(
'\t')
tids, tinseqs, tmqs1, talns1, tmqs2, talns2, tpinds, tcinds = mergedBedpeInfo[
key].split('\t')
# the investigated key is sufficiently far from the current line in the input file and no additional line to merge is expected. therefore flush the key and information
if F[0] != tchr1 or int(F[1]) > int(tend1) + check_margin_size:
# obtain the most frequent junction
junc_counter = collections.Counter(
mergedJunction[key].split(';'))
best_junc = junc_counter.most_common(1)[0][0]
btchr1, btend1, btdir1, btchr2, btend2, btdir2, btinseq = best_junc.split(
',')
btstart1 = str(int(btend1) - 1)
btstart2 = str(int(btend2) - 1)
print >> hOUT, '\t'.join([btchr1, btstart1, btend1, btchr2, btstart2, btend2, \
tids, btinseq, btdir1, btdir2, tmqs1, talns1, \
tmqs2, talns2, tpinds, tcinds]) + '\t' + \
mergedJunction[key]
# add to the deletion list (later the key will removed from the dictionaries)
delList.append(key)
continue
else:
# check whether the investigated key and the current line should be merged or not
if F[0] == tchr1 and F[3] == tchr2 and F[8] == tdir1 and F[
9] == tdir2:
flag = 0
# detailed check on the junction position considering inserted sequences
if F[8] == "+":
expectedDiffSize = (int(F[2]) - int(tend1)) + (
len(F[7]) - int(inseqSize))
if (F[9] == "+" and int(F[5])
== int(tend2) - int(expectedDiffSize)) or (
F[9] == "-" and int(F[5])
== int(tend2) + int(expectedDiffSize)):
flag = 1
else:
expectedDiffSize = (int(F[2]) - int(tend1)) + (
int(inseqSize) - len(F[7]))
if (F[9] == "+" and int(F[5])
== int(tend2) + int(expectedDiffSize)) or (
F[9] == "-" and int(F[5])
== int(tend2) - int(expectedDiffSize)):
flag = 1
# if the junction position and direciton match
if flag == 1:
match = 1
newIds = tids + ';' + F[6]
newInseqs = tinseqs + ';' + F[7]
newMqs1 = tmqs1 + ';' + F[10]
newAlns1 = talns1 + ';' + F[11]
newMqs2 = tmqs2 + ';' + F[12]
newAlns2 = talns2 + ';' + F[13]
newPinds = tpinds + ';' + F[14]
newCinds = tcinds + ';' + F[15]
mergedBedpeInfo[key] = '\t'.join([
newIds, newInseqs, newMqs1, newAlns1, newMqs2,
newAlns2, newPinds, newCinds
])
# check whether the inserted sequence should be reverse-complemented
tinseq = F[7]
if F[7] != "---" and F[8] == F[9] and F[15] == "2":
# tinseq = str(Bio.Seq.Seq(F[7]).reverse_complement())
tinseq = utils.reverseComplement(F[7])
mergedJunction[
key] = mergedJunction[key] + ";" + ','.join(
[F[0], F[2], F[8], F[3], F[5], F[9], tinseq])
for item in delList:
del mergedBedpeInfo[item]
del mergedJunction[item]
# if the current line in the input file does not match any of the pooled keys
if match == 0:
newKey = '\t'.join([
F[0], F[1], F[2], F[3], F[4], F[5], F[8], F[9],
str(len(F[7]))
])
mergedBedpeInfo[newKey] = F[6] + '\t' + F[7] + '\t' + '\t'.join(
F[10:16])
# check whether the inserted sequence should be reverse-complemented
tinseq = F[7]
if F[7] != "---" and F[8] == F[9] and F[15] == "2":
# tinseq = str(Bio.Seq.Seq(F[7]).reverse_complement())
tinseq = utils.reverseComplement(F[7])
mergedJunction[newKey] = ','.join(
[F[0], F[2], F[8], F[3], F[5], F[9], tinseq])
hIN.close()
# last treatment
for key in sorted(mergedBedpeInfo):
tchr1, tstart1, tend1, tchr2, tstart2, tend2, tdir1, tdir2, inseqSize = key.split(
'\t')
tids, tinseqs, tmqs1, talns1, tmqs2, talns2, tpinds, tcinds = mergedBedpeInfo[
key].split('\t')
# obtain the most frequent junction
junc_counter = collections.Counter(mergedJunction[key].split(';'))
best_junc = junc_counter.most_common(1)[0][0]
btchr1, btend1, btdir1, btchr2, btend2, btdir2, btinseq = best_junc.split(
',')
btstart1 = str(int(btend1) - 1)
btstart2 = str(int(btend2) - 1)
print >> hOUT, '\t'.join([btchr1, btstart1, btend1, btchr2, btstart2, btend2, \
tids, btinseq, btdir1, btdir2, tmqs1, talns1, \
tmqs2, talns2, tpinds, tcinds]) + '\t' + \
mergedJunction[key]
hOUT.close()
def getRefAltForSV(outputFilePath, juncChr1, juncPos1, juncDir1, juncChr2,
juncPos2, juncDir2, juncSeq, reference_genome,
split_refernece_thres, validate_sequence_length):
"""
for short SV (mid-range (<= split_refernece_thres bp) deletion, tandem duplication), we get the two sequence
for large SV (> split_refernece_thres bp), we get three sequence (one joint sequence by two break points, and two reference sequences around the break points)
the only concern is short inversion... (are there some somatic short inversion?)
however, this will be filtered beforehand by the "cover filter", and maybe we have to give up detecting these class of SVs.
"""
hOUT = open(outputFilePath, 'w')
if juncSeq == "---": juncSeq = ""
# for mid-range deletion or tandem duplication
if juncChr1 == juncChr2 and abs(int(juncPos1) - int(
juncPos2)) <= split_refernece_thres and juncDir1 != juncDir2:
seq = ""
for item in pysam.faidx(
reference_genome, juncChr1 + ":" +
str(int(juncPos1) - validate_sequence_length) + "-" +
str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([
juncChr1,
str(juncPos1), juncDir1, juncChr2,
str(juncPos2), juncDir2
]) + "_ref"
print >> hOUT, seq
# for mid-range deletion
if juncDir1 == "+" and juncDir2 == "-":
seq = ""
for item in pysam.faidx(
reference_genome, juncChr1 + ":" +
str(int(juncPos1) - validate_sequence_length) + "-" +
str(juncPos1)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
seq = seq + juncSeq
for item in pysam.faidx(
reference_genome, juncChr2 + ":" + str(juncPos2) + "-" +
str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([
juncChr1,
str(juncPos1), juncDir1, juncChr2,
str(juncPos2), juncDir2
]) + "_alt"
print >> hOUT, seq
# for mid-range tandem duplication
else:
seq = ""
for item in pysam.faidx(
reference_genome, juncChr2 + ":" +
str(int(juncPos2) - validate_sequence_length) + "-" +
str(juncPos2)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
seq = seq + juncSeq
for item in pysam.faidx(
reference_genome, juncChr1 + ":" + str(juncPos1) + "-" +
str(int(juncPos1) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([
juncChr1,
str(juncPos1), juncDir1, juncChr2,
str(juncPos2), juncDir2
]) + "_alt"
print >> hOUT, seq
else:
seq = ""
for item in pysam.faidx(
reference_genome, juncChr1 + ":" +
str(int(juncPos1) - validate_sequence_length) + "-" +
str(int(juncPos1) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([
juncChr1,
str(juncPos1), juncDir1, juncChr2,
str(juncPos2), juncDir2
]) + "_ref1"
print >> hOUT, seq
seq = ""
for item in pysam.faidx(
reference_genome, juncChr2 + ":" +
str(int(juncPos2) - validate_sequence_length) + "-" +
str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([
juncChr1,
str(juncPos1), juncDir1, juncChr2,
str(juncPos2), juncDir2
]) + "_ref2"
print >> hOUT, seq
seq = ""
if juncDir1 == "+":
tseq = ""
for item in pysam.faidx(
reference_genome, juncChr1 + ":" +
str(int(juncPos1) - validate_sequence_length) + "-" +
str(juncPos1)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
else:
tseq = ""
for item in pysam.faidx(
reference_genome, juncChr1 + ":" + str(juncPos1) + "-" +
str(int(juncPos1) + validate_sequence_length)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
tseq = utils.reverseComplement(tseq)
seq = tseq + juncSeq
if juncDir2 == "-":
tseq = ""
for item in pysam.faidx(
reference_genome, juncChr2 + ":" + str(juncPos2) + "-" +
str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
else:
tseq = ""
for item in pysam.faidx(
reference_genome, juncChr2 + ":" +
str(int(juncPos2) - validate_sequence_length) + "-" +
str(juncPos2)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
tseq = utils.reverseComplement(tseq)
seq = seq + tseq
print >> hOUT, '>' + ','.join([
juncChr1,
str(juncPos1), juncDir1, juncChr2,
str(juncPos2), juncDir2
]) + "_alt"
print >> hOUT, seq
hOUT.close()
def extractSVReadPairs(bamFilePath, outputFilePath, juncChr1, juncPos1, juncDir1, juncChr2, juncPos2, juncDir2, max_depth, search_length, search_margin):
"""
read pairs containing break points are extracted. (yshira 2015/04/23)
The exact condition is as follows:
1. one of the read in the pair has the break point of the SV candidate
2. the start positions of the read pairs are within 800bp of the break point of the SV candidate
Some minor concern for the above conditions are:
1. Depending on the choice of the "start position" or "end position", the distance between the read and break point differs. This can generate slight bias...
(but I believe we can recover this by setting sufficient margin (800bp), and summarize the alignment result carefully.)
2. Maybe, for some of the read pair, the result of alignment is obvious. But should we re-align them?
"""
bamfile = pysam.Samfile(bamFilePath, 'rb')
# if the #sequence read is over the `maxDepth`, then that key is ignored
depthFlag = 0
if bamfile.count(juncChr1, int(juncPos1) - 1, int(juncPos1) + 1) >= max_depth: depthFlag = 1
if bamfile.count(juncChr2, int(juncPos2) - 1, int(juncPos2) + 1) >= max_depth: depthFlag = 1
if depthFlag == 1:
print >> sys.stderr, "sequence depth exceeds the threshould for: " + ','.join([juncChr1, juncPos1, juncDir1, juncChr2, juncPos2, juncDir2])
return 1
hOUT = open(outputFilePath, 'w')
readID2exist = {}
for read in bamfile.fetch(juncChr1, max(0, int(juncPos1) - search_length), int(juncPos1) + search_length):
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
chr_current = bamfile.getrname(read.tid)
pos_current = int(read.pos + 1)
dir_current = ("-" if flags[4] == "1" else "+")
chr_pair = bamfile.getrname(read.rnext)
pos_pair = int(read.pnext + 1)
dir_pair = ("-" if flags[5] == "1" else "+")
# the read (with margin) contains break point
if pos_current - search_margin <= int(juncPos1) <= (read.aend - 1) + search_margin:
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr1 and pos_current <= int(juncPos1) <= pos_pair and dir_current == "+" and dir_pair == "-":
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr2:
juncFlag = 0
if juncDir1 == "+" and juncDir2 == "+" and pos_current <= int(juncPos1) and pos_pair <= int(juncPos2): juncFlag = 1
if juncDir1 == "+" and juncDir2 == "-" and pos_current <= int(juncPos1) and pos_pair >= int(juncPos2): juncFlag = 1
if juncDir1 == "-" and juncDir2 == "+" and pos_current >= int(juncPos1) and pos_pair <= int(juncPos2): juncFlag = 1
if juncDir1 == "-" and juncDir2 == "-" and pos_current >= int(juncPos1) and pos_pair >= int(juncPos2): juncFlag = 1
if juncFlag == 1:
readID2exist[read.qname] = 1
for read in bamfile.fetch(juncChr2, max(0, int(juncPos2) - search_length), int(juncPos2) + search_length):
if read.qname == "ST-E00104:162:H03UUALXX:5:1222:21168:16006":
pass
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
chr_current = bamfile.getrname(read.tid)
pos_current = int(read.pos + 1)
dir_current = ("-" if flags[4] == "1" else "+")
chr_pair = bamfile.getrname(read.rnext)
pos_pair = int(read.pnext + 1)
dir_pair = ("-" if flags[5] == "1" else "+")
# the read (with margin) contains break point
if pos_current - search_margin <= int(juncPos2) <= (read.aend - 1) + search_margin:
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr2 and pos_current <= int(juncPos2) <= pos_pair and dir_current == "+" and dir_pair == "-":
readID2exist[read.qname] = 1
# the read pair covers break point
if chr_pair == juncChr1:
juncFlag = 0
if juncDir2 == "+" and juncDir1 == "+" and pos_current <= int(juncPos2) and pos_pair <= int(juncPos1): juncFlag = 1
if juncDir2 == "+" and juncDir1 == "-" and pos_current <= int(juncPos2) and pos_pair >= int(juncPos1): juncFlag = 1
if juncDir2 == "-" and juncDir1 == "+" and pos_current >= int(juncPos2) and pos_pair <= int(juncPos1): juncFlag = 1
if juncDir2 == "-" and juncDir1 == "-" and pos_current >= int(juncPos2) and pos_pair >= int(juncPos1): juncFlag = 1
if juncFlag == 1:
readID2exist[read.qname] = 1
readID2seq1 = {}
readID2seq2 = {}
complement = {'A': 'T', 'C': 'G', 'G': 'C', 'T': 'A', 'N': 'N'}
for read in bamfile.fetch(juncChr1, max(0, int(juncPos1) - search_length), int(juncPos1) + search_length):
if read.qname in readID2exist:
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
tempSeq = ""
if flags[4] == "1":
tempSeq = utils.reverseComplement(str(read.seq))
else:
tempSeq = read.seq
# the first read
if flags[6] == "1":
readID2seq1[read.qname] = tempSeq
else:
readID2seq2[read.qname] = tempSeq
for read in bamfile.fetch(juncChr2, max(0, int(juncPos2) - search_length), int(juncPos2) + search_length):
if read.qname in readID2exist:
# get the flag information
flags = format(int(read.flag), "#014b")[:1:-1]
# skip unmapped read
if flags[2] == "1" or flags[3] == "1": continue
# skip supplementary alignment
if flags[8] == "1" or flags[11] == "1": continue
# skip duplicated reads
if flags[10] == "1": continue
tempSeq = ""
if flags[4] == "1":
tempSeq = utils.reverseComplement(str(read.seq))
else:
tempSeq = read.seq
# the first read
if flags[6] == "1":
readID2seq1[read.qname] = tempSeq
else:
readID2seq2[read.qname] = tempSeq
for readID in readID2seq1:
if readID in readID2seq2:
print >> hOUT, '>' + readID + '/1'
print >> hOUT, readID2seq1[readID]
print >> hOUT, '>' + readID + '/2'
print >> hOUT, readID2seq2[readID]
bamfile.close()
hOUT.close()
return 0
def getRefAltForSV(outputFilePath, juncChr1, juncPos1, juncDir1, juncChr2, juncPos2, juncDir2, juncSeq, reference_genome, split_refernece_thres, validate_sequence_length):
"""
for short SV (mid-range (<= split_refernece_thres bp) deletion, tandem duplication), we get the two sequence
for large SV (> split_refernece_thres bp), we get three sequence (one joint sequence by two break points, and two reference sequences around the break points)
the only concern is short inversion... (are there some somatic short inversion?)
however, this will be filtered beforehand by the "cover filter", and maybe we have to give up detecting these class of SVs.
"""
hOUT = open(outputFilePath, 'w')
if juncSeq == "---": juncSeq = ""
# for mid-range deletion or tandem duplication
if juncChr1 == juncChr2 and abs(int(juncPos1) - int(juncPos2)) <= split_refernece_thres and juncDir1 != juncDir2:
seq = ""
for item in pysam.faidx(reference_genome, juncChr1 + ":" + str(int(juncPos1) - validate_sequence_length) + "-" + str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([juncChr1, str(juncPos1), juncDir1, juncChr2, str(juncPos2), juncDir2]) + "_ref"
print >> hOUT, seq
# for mid-range deletion
if juncDir1 == "+" and juncDir2 == "-":
seq = ""
for item in pysam.faidx(reference_genome, juncChr1 + ":" + str(int(juncPos1) - validate_sequence_length) + "-" + str(juncPos1)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
seq = seq + juncSeq
for item in pysam.faidx(reference_genome, juncChr2 + ":" + str(juncPos2) + "-" + str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([juncChr1, str(juncPos1), juncDir1, juncChr2, str(juncPos2), juncDir2]) + "_alt"
print >> hOUT, seq
# for mid-range tandem duplication
else:
seq = ""
for item in pysam.faidx(reference_genome, juncChr2 + ":" + str(int(juncPos2) - validate_sequence_length) + "-" + str(juncPos2)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
seq = seq + juncSeq
for item in pysam.faidx(reference_genome, juncChr1 + ":" + str(juncPos1) + "-" + str(int(juncPos1) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([juncChr1, str(juncPos1), juncDir1, juncChr2, str(juncPos2), juncDir2]) + "_alt"
print >> hOUT, seq
else:
seq = ""
for item in pysam.faidx(reference_genome, juncChr1 + ":" + str(int(juncPos1) - validate_sequence_length) + "-" + str(int(juncPos1) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([juncChr1, str(juncPos1), juncDir1, juncChr2, str(juncPos2), juncDir2]) + "_ref1"
print >> hOUT, seq
seq = ""
for item in pysam.faidx(reference_genome, juncChr2 + ":" + str(int(juncPos2) - validate_sequence_length) + "-" + str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
seq = seq + item.rstrip('\n').upper()
print >> hOUT, '>' + ','.join([juncChr1, str(juncPos1), juncDir1, juncChr2, str(juncPos2), juncDir2]) + "_ref2"
print >> hOUT, seq
seq = ""
if juncDir1 == "+":
tseq = ""
for item in pysam.faidx(reference_genome, juncChr1 + ":" + str(int(juncPos1) - validate_sequence_length) + "-" + str(juncPos1)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
else:
tseq = ""
for item in pysam.faidx(reference_genome, juncChr1 + ":" + str(juncPos1) + "-" + str(int(juncPos1) + validate_sequence_length)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
tseq = utils.reverseComplement(tseq)
seq = tseq + juncSeq
if juncDir2 == "-":
tseq = ""
for item in pysam.faidx(reference_genome, juncChr2 + ":" + str(juncPos2) + "-" + str(int(juncPos2) + validate_sequence_length)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
else:
tseq = ""
for item in pysam.faidx(reference_genome, juncChr2 + ":" + str(int(juncPos2) - validate_sequence_length) + "-" + str(juncPos2)):
if item[0] == ">": continue
tseq = tseq + item.rstrip('\n').upper()
tseq = utils.reverseComplement(tseq)
seq = seq + tseq
print >> hOUT, '>' + ','.join([juncChr1, str(juncPos1), juncDir1, juncChr2, str(juncPos2), juncDir2]) + "_alt"
print >> hOUT, seq
hOUT.close()
def clusterJunction(inputFilePath, outputFilePath, check_margin_size):
"""
script for merging and summarizing junction read pairs
"""
hIN = open(inputFilePath, 'r')
hOUT = open(outputFilePath, 'w')
mergedBedpeInfo = {}
mergedJunction = {}
for line in hIN:
F = line.rstrip('\n').split('\t')
match = 0
delList = []
for key in sorted(mergedBedpeInfo):
tchr1, tstart1, tend1, tchr2, tstart2, tend2, tdir1, tdir2, inseqSize = key.split('\t')
tids, tinseqs, tmqs1, talns1, tmqs2, talns2, tpinds, tcinds = mergedBedpeInfo[key].split('\t')
# the investigated key is sufficiently far from the current line in the input file and no additional line to merge is expected. therefore flush the key and information
if F[0] != tchr1 or int(F[1]) > int(tend1) + check_margin_size:
# obtain the most frequent junction
junc_counter = collections.Counter(mergedJunction[key].split(';'))
best_junc = junc_counter.most_common(1)[0][0]
btchr1, btend1, btdir1, btchr2, btend2, btdir2, btinseq = best_junc.split(',')
btstart1 = str(int(btend1) - 1)
btstart2 = str(int(btend2) - 1)
print >> hOUT, '\t'.join([btchr1, btstart1, btend1, btchr2, btstart2, btend2, \
tids, btinseq, btdir1, btdir2, tmqs1, talns1, \
tmqs2, talns2, tpinds, tcinds]) + '\t' + \
mergedJunction[key]
# add to the deletion list (later the key will removed from the dictionaries)
delList.append(key)
continue
else:
# check whether the investigated key and the current line should be merged or not
if F[0] == tchr1 and F[3] == tchr2 and F[8] == tdir1 and F[9] == tdir2:
flag = 0
# detailed check on the junction position considering inserted sequences
if F[8] == "+":
expectedDiffSize = (int(F[2]) - int(tend1)) + (len(F[7]) - int(inseqSize))
if (F[9] == "+" and int(F[5]) == int(tend2) - int(expectedDiffSize)) or (F[9] == "-" and int(F[5]) == int(tend2) + int(expectedDiffSize)):
flag = 1
else:
expectedDiffSize = (int(F[2]) - int(tend1)) + (int(inseqSize) - len(F[7]))
if (F[9] == "+" and int(F[5]) == int(tend2) + int(expectedDiffSize)) or (F[9] == "-" and int(F[5]) == int(tend2) - int(expectedDiffSize)):
flag = 1
# if the junction position and direciton match
if flag == 1:
match = 1
newIds = tids + ';' + F[6]
newInseqs = tinseqs + ';' + F[7]
newMqs1 = tmqs1 + ';' + F[10]
newAlns1 = talns1 + ';' + F[11]
newMqs2 = tmqs2 + ';' + F[12]
newAlns2 = talns2 + ';' + F[13]
newPinds = tpinds + ';' + F[14]
newCinds = tcinds + ';' + F[15]
mergedBedpeInfo[key] = '\t'.join([newIds, newInseqs, newMqs1, newAlns1, newMqs2, newAlns2, newPinds, newCinds])
# check whether the inserted sequence should be reverse-complemented
tinseq = F[7]
if F[7] != "---" and F[8] == F[9] and F[15] == "2":
# tinseq = str(Bio.Seq.Seq(F[7]).reverse_complement())
tinseq = utils.reverseComplement(F[7])
mergedJunction[key] = mergedJunction[key] + ";" + ','.join([F[0], F[2], F[8], F[3], F[5], F[9], tinseq])
for item in delList:
del mergedBedpeInfo[item]
del mergedJunction[item]
# if the current line in the input file does not match any of the pooled keys
if match == 0:
newKey = '\t'.join([F[0], F[1], F[2], F[3], F[4], F[5], F[8], F[9], str(len(F[7]))])
mergedBedpeInfo[newKey] = F[6] + '\t' + F[7] + '\t' + '\t'.join(F[10:16])
# check whether the inserted sequence should be reverse-complemented
tinseq = F[7]
if F[7] != "---" and F[8] == F[9] and F[15] == "2":
# tinseq = str(Bio.Seq.Seq(F[7]).reverse_complement())
tinseq = utils.reverseComplement(F[7])
mergedJunction[newKey] = ','.join([F[0], F[2], F[8], F[3], F[5], F[9], tinseq])
hIN.close()
# last treatment
for key in sorted(mergedBedpeInfo):
tchr1, tstart1, tend1, tchr2, tstart2, tend2, tdir1, tdir2, inseqSize = key.split('\t')
tids, tinseqs, tmqs1, talns1, tmqs2, talns2, tpinds, tcinds = mergedBedpeInfo[key].split('\t')
# obtain the most frequent junction
junc_counter = collections.Counter(mergedJunction[key].split(';'))
best_junc = junc_counter.most_common(1)[0][0]
btchr1, btend1, btdir1, btchr2, btend2, btdir2, btinseq = best_junc.split(',')
btstart1 = str(int(btend1) - 1)
btstart2 = str(int(btend2) - 1)
print >> hOUT, '\t'.join([btchr1, btstart1, btend1, btchr2, btstart2, btend2, \
tids, btinseq, btdir1, btdir2, tmqs1, talns1, \
tmqs2, talns2, tpinds, tcinds]) + '\t' + \
mergedJunction[key]
hOUT.close()
'TTA': 'L',
'TTG': 'L',
'TAC': 'Y',
'TAT': 'Y',
'TAA': '*',
'TAG': '*',
'TGC': 'C',
'TGT': 'C',
'TGA': '*',
'TGG': 'W',
}
for header, sequence in utils.parse_fasta(sequenceFile):
positiveStrand = ""
longestCDS = 0
strands = [sequence, utils.reverseComplement(sequence)]
for strand in strands:
for frame in range(3):
proteinSequence = ""
for fragment in range(frame, len(strand), 3):
codon = strand[fragment:fragment + 3]
if len(codon) != 3:
continue
try:
proteinSequence += codon2aminoacid[codon]
except KeyError:
proteinSequence += 'X'
matches = regex_orf.findall(proteinSequence)
allORFs = "".join([x for x in matches if x])
if len(allORFs) / float(len(strand)) > longestCDS:
longestCDS = len(allORFs) / float(len(strand))