def generate_read_from_sam(input_lines, keep_seq=True):
"""Convert a set of lines from a SAM file into an RNASeqRead object"""
generated_read = RNASeqRead(ID=None,
Nmap=0,
primary={},
secondary={},
seq1=None,
seq2=None)
for line in input_lines:
l = line.rstrip().split('\t')
ID = l[0]
if not generated_read.ID:
generated_read.ID = ID
assert ID == generated_read.ID, 'ERROR: nonmatching IDs in input_lines:\n{}\t{}'.format(
generated_read.ID, ID)
attributes = dict([(':'.join(i.split(':')[0:2]), i.split(':')[-1])
for i in l[11:]])
# Nmap = int(attributes['NH:i'])
# if generated_read.Nmap == 0:
# generated_read.Nmap = Nmap
# assert Nmap == generated_read.Nmap, 'ERROR: inconsistent Nmap score for {}'.format(ID)
SAMflags = bin(int(l[1]))[2:]
SAMflags = '0' * (12 - len(SAMflags)) + SAMflags
# Interpret the binary SAM flags
is_paired = bool(int(SAMflags[-1]))
pair_is_mapped = bool(int(SAMflags[-2]))
read_reverse = bool(int(SAMflags[-5]))
first_in_pair = bool(int(SAMflags[-7]))
secondary = bool(int(SAMflags[-9]))
supplementary = bool(int(SAMflags[-12]))
# read_unmapped = bool(int(SAMflags[-3]))
# mate_unmapped = bool(int(SAMflags[-4]))
# mate_reverse = bool(int(SAMflags[-6]))
# second_in_pair = bool(int(SAMflags[-8]))
# quality_fail = bool(int(SAMflags[-10]))
# pcr_duplicate = bool(int(SAMflags[-11]))
###
chrom = l[2]
pos = int(l[3])
pairpos = int(l[7])
mapscore = int(l[4])
cigar = l[5]
seq = l[9]
generated_read.paired = is_paired
if keep_seq:
# store the + stranded read of the respective mate pair
if read_reverse:
rcseq = fu.rc(seq)
if first_in_pair or not is_paired:
# store reverse complement in seq1
if not generated_read.seq1:
generated_read.seq1 = rcseq
assert rcseq == generated_read.seq1, 'ERROR: nonmatching sequence in input_lines:\n{}\t{}'.format(
generated_read.seq1, rcseq)
else:
# store reverse complement in seq2
if not generated_read.seq2:
generated_read.seq2 = rcseq
assert rcseq == generated_read.seq2, 'ERROR: nonmatching sequence in input_lines:\n{}\t{}'.format(
generated_read.seq2, rcseq)
else:
if first_in_pair or not is_paired:
# store in seq1
if not generated_read.seq1:
generated_read.seq1 = seq
assert seq == generated_read.seq1, 'ERROR: nonmatching sequence in input_lines:\n{}\t{}'.format(
generated_read.seq1, seq)
else:
# store in seq2
if not generated_read.seq2:
generated_read.seq2 = seq
assert seq == generated_read.seq2, 'ERROR: nonmatching sequence in input_lines:\n{}\t{}'.format(
generated_read.seq2, seq)
if read_reverse:
if first_in_pair or not is_paired:
strand = '-'
else:
strand = '+'
else:
if first_in_pair or not is_paired:
strand = '+'
else:
strand = '-'
if first_in_pair or not is_paired:
if secondary or supplementary:
if pos in generated_read.secondary:
generated_read.secondary[pos][2].append((pos, cigar, 1))
else:
generated_read.secondary[pos] = [
chrom, strand, [(pos, cigar, 1)]
]
else:
if pos in generated_read.primary:
generated_read.primary[pos][2].append((pos, cigar, 1))
else:
generated_read.primary[pos] = [
chrom, strand, [(pos, cigar, 1)]
]
else:
if not pair_is_mapped:
continue
if secondary or supplementary:
if pairpos in generated_read.secondary:
generated_read.secondary[pairpos][2].append(
(pos, cigar, 2))
else:
generated_read.secondary[pairpos] = [
chrom, strand, [(pos, cigar, 2)]
]
else:
if pairpos in generated_read.primary:
generated_read.primary[pairpos][2].append((pos, cigar, 2))
else:
generated_read.primary[pairpos] = [
chrom, strand, [(pos, cigar, 2)]
]
return generated_read
s = '+'
for t in group:
positions = set(
flatten([
list(range(a - 1, b))
for a, b in zip(transcripts[t].get_exon_start(),
transcripts[t].get_exon_end())
]))
if args.FEATURE != 'transcript':
positions = sorted(list(positions))
strand = transcripts[t].strand
sequence = ''.join([genome[chrom][i] for i in positions])
if strand == '-':
aa, ss, f = fu.longest_orf(fu.rc(sequence))
if len(aa) < MIN_ORF:
continue
ORFstart, ORFstop = ss
if ORFstart == 0:
ORFstart = 1
if ORFstop == len(positions):
ORFstop = len(positions) - 1
if args.FEATURE == '5UTR':
positions = positions[-ORFstart:]
elif args.FEATURE == 'CDS':
positions = positions[-ORFstop:-ORFstart]
elif args.FEATURE == '3UTR':
positions = positions[:-ORFstop]
else:
transcript_coverage = np.empty(0, dtype='float32')
if args.write_fasta:
transcript_fasta = ''
if strand not in ['+', '-']:
print('# WARNING: {} is unstranded'.format(ID))
continue
for a, b in zip(exon_starts, exon_ends):
exon = coverage[strand][chrom][a - 1:b]
transcript_coverage = np.append(transcript_coverage, exon)
if args.write_fasta:
transcript_fasta += genome[chrom][a - 1:b]
if strand == '-': # Flip the strand for minus-stranded data
transcript_coverage = transcript_coverage[::-1]
if args.write_fasta:
transcript_fasta = fu.rc(transcript_fasta)
for line in vector_to_bedgraph_lines(transcript_coverage, ID):
output_file.write(line + '\n')
if args.write_fasta:
fasta_outfile.write('>{}\n'.format(ID))
fasta_outfile.write('{}\n'.format(transcript_fasta))
output_file.close()
if args.write_fasta:
fasta_outfile.close()
feature, len(bed_features[feature]['positions'])))
values = [
sum([
coverage[graph][bed_features[feature]['chrom']].get(i, 0)
for i in bed_features[feature]['positions']
]) for graph in ingraphs
]
if args.g_content:
nucleotides = [
genome[bed_features[feature]['chrom']][i]
for i in bed_features[feature]['positions']
]
if bed_features[feature]['strand'] == '-':
nucleotides = [i for i in fu.rc(''.join(nucleotides))]
G = round(
float(sum([i == 'G' for i in nucleotides])) / len(nucleotides), 3)
g_file.write('{}\t{}\n'.format(feature, G))
if args.BED_OUT:
outfile.write('\t'.join([
str(i) for i in [
bed_features[feature]['chrom'],
min(bed_features[feature]['positions']),
max(bed_features[feature]['positions']),
feature,
'.',
bed_features[feature]['strand'],
] + values
]) + '\n')
# args.reference_GFF
# ))
#################################
### EVALUATE THE GIVEN SAMPLE ###
#################################
for ID in ref_IDs:
input_transcript = ref_transcripts[ID]
chrom = input_transcript.chrom
strand = input_transcript.strand
# Gather information about exons
exon_starts = input_transcript.get_exon_start()
exon_ends = input_transcript.get_exon_end()
if len(exon_starts) != len(exon_ends):
print('# WARNING: inconsistent exon structure with {}'.format(ID))
continue
# Make a list of all nucleotide positions in the input transcript
positions = flatten(
[list(range(a - 1, b)) for a, b in zip(exon_starts, exon_ends)])
length = len(positions)
positions = [i for i in positions if i < chromosomes[chrom] and i > 0]
# Get transcript's nucleotide sequence from the FASTA file
sequence = ''.join([genome[chrom][i] for i in positions])
if strand == '-':
sequence = fu.rc(sequence)
print('>{}\n{}'.format(ID, sequence))