def align_consensus(cons_file_1, cons_file_2):
""" Align consensus.faa to each other and to the HIV reference
"""
from pythonlib import EmbossStandalone
from pythonlib.MarkxIO import Markx10Iterator
needle_exe = 'needle'
out_file = 'map_cons.needle'
EmbossStandalone.needle(needle_exe,
cons_file_1,
cons_file_2,
out=out_file,
aglobal3='False')
alignment = Markx10Iterator(open(out_file)).next()
cons_1 = alignment.get_seq_by_num(0).tostring().upper()
cons_2 = alignment.get_seq_by_num(1).tostring().upper()
map = []
for c in zip(cons_1, cons_2):
assert not (c[0] == c[1] and c[1] == '-'), 'not two gaps'
if c[0] == '-':
map.append(1)
elif c[1] == '-':
map.append(2)
else:
map.append(0)
return map
length += float(len_seq)
length2 += float(len_seq * len_seq)
# readdict[read.] = [seq,len_seq]
n += 1.
meanlr = length / n
stdlr = math.sqrt((n * length2 - length * length) / (n * n - n))
allowed_length = [meanlr - acclength * stdlr, meanlr + (1 + acclength) * stdlr]
print >> sys.stderr, 'Allowed interval for length is', allowed_length
if not os.path.isfile('tmp_align_f.needle'):
print >> sys.stderr, 'needle forward'
EmbossStandalone.needle(needle_exe,
options.ref,
f_fasta_forward_filename,
out='tmp_align_f.needle',
gapopen=6.0,
gapext=3.0,
aglobal3='False')
"""
else:
print >>sys.stderr, 'The alignment file tmp_align_f.needle is already present'
statinfo = os.stat('tmp_align_f.needle')
age_sec = time.time() - statinfo.st_mtime
if age_sec > 3600:
print >>sys.stderr, 'Warning: it was modified more than an hour ago'
age = time.gmtime(age_sec)
print >>sys.stderr, 'If you want to run the alignment again, remove it'
print >>sys.stderr, "using existing 'tmp_align_f.needle'..."
"""
def align_reads(filename):
"""reads the file with reads, align them with the reference,
returns a dictionary with reads (in-dels are discarded)
and starting position with respect to the reference
"""
from pythonlib import EmbossStandalone
from pythonlib.MarkxIO import Markx10Iterator
needle_exe = 'needle'
aligned_reads = {}
f_fasta = open(filename)
seqlist = list(SeqIO.parse(f_fasta, 'fasta'))
countreads = len(seqlist)
# forward...
f_fasta_forward_filename = 'tmp_reads_f.fas'
f_fasta_forward = open(f_fasta_forward_filename, 'w')
SeqIO.write(seqlist, f_fasta_forward, 'fasta')
f_fasta_forward.close()
# ...and reverse
for seq in seqlist:
seq.seq = seq.seq.reverse_complement()
f_fasta.close()
f_fasta_reverse_filename = 'tmp_reads_r.fas'
f_fasta_reverse = open(f_fasta_reverse_filename, 'w')
SeqIO.write(seqlist, f_fasta_reverse, 'fasta')
f_fasta_reverse.close()
print >> sys.stderr, 'Found', countreads, 'reads'
if not os.path.isfile('tmp_align_f.needle'):
print >> sys.stderr, 'needle forward'
EmbossStandalone.needle(needle_exe,
ref_genome,
f_fasta_forward_filename,
out='tmp_align_f.needle',
gapopen=6.0,
gapext=3.0,
aglobal3='False',
adesshow3='True')
if not os.path.isfile('tmp_align_r.needle'):
print >> sys.stderr, 'needle backward'
EmbossStandalone.needle(needle_exe,
ref_genome,
f_fasta_reverse_filename,
out='tmp_align_r.needle',
gapopen=6.0,
gapext=3.0,
aglobal3='False',
adesshow3='True')
f_forward = open('tmp_align_f.needle')
f_reverse = open('tmp_align_r.needle')
forwardaligniter = Markx10Iterator(f_forward)
reversealigniter = Markx10Iterator(f_reverse)
count_forward = 0
count_reverse = 0
while True:
# pos += 1
# print >> sys.stderr, '\x1B[1A\x1B[2K', pos
try:
f_align = forwardaligniter.next()
r_align = reversealigniter.next()
except:
break
if f_align is None or r_align is None:
break
assert f_align.get_all_seqs()[1].id == r_align.get_all_seqs(
)[1].id, 'same seq back and forward'
this_id = f_align.get_all_seqs()[1].id
if float(f_align._annotations['sw_score']) > float(
r_align._annotations['sw_score']):
tmp = f_align.get_seq_by_num(1).tostring().upper()
refseq = f_align.get_seq_by_num(0).tostring().upper()
count_forward += 1
else:
tmp = r_align.get_seq_by_num(1).tostring().upper()
refseq = r_align.get_seq_by_num(0).tostring().upper()
count_reverse += 1
q_align_start = len(tmp) - len(tmp.lstrip('-'))
q_align_end = len(tmp.rstrip('-'))
m_align_start = len(refseq) - len(refseq.lstrip('-'))
m_align_end = len(refseq.rstrip('-'))
align_start = max(m_align_start, q_align_start)
align_end = min(m_align_end, q_align_end)
this_read = []
for c in zip(refseq[align_start:align_end + 1],
tmp[align_start:align_end + 1]):
if c[0] != '-' and c[1] != '-':
this_read.append(c[1])
elif c[1] == '-':
this_read.append(c[0])
elif c[0] == '-':
pass
aligned_reads[this_id] = [''.join(this_read), align_start]
return aligned_reads
def align_to_ref(al_exe, ref_file, reads_file, gen_length):
"""
Calls water standalone program to align reads to reference genome
"""
from pythonlib import EmbossStandalone
import MyAlignIO
import time
max_read_length = 300
format = 'markx10'
align_file = '%s.needle' % reads_file.rstrip('.fas')
out_reads = {}
cov_prof = [0] * (2 * gen_length + max_read_length)
if not os.path.isfile(align_file):
print 'Aligning reads via Needleman-Wunsch algorithm'
EmbossStandalone.needle(al_exe,
ref_file,
reads_file,
out=align_file,
gapopen=go_default,
gapext=ge_default,
aglobal3='False')
else:
print 'The alignment file', align_file, 'is already present'
statinfo = os.stat(align_file)
age_sec = time.time() - statinfo.st_mtime
if age_sec > 3600:
print 'Warning: it was modified more than an hour ago'
age = time.gmtime(age_sec)
print 'If you want to run the alignment again, remove it'
assert os.path.isfile(align_file), 'File %s not found' % align_file
handle = open(align_file, 'rU')
print 'Parsing alignment output'
for alin in MyAlignIO.parse(handle, format):
assert len(alin.get_all_seqs()) == 2, "Should be pairwise!"
alength = int(alin.get_alignment_length())
# print 'Alignment is', alength, 'bases long'
record = iter(alin)
# These are the information of the query sequence, i.e. the reference
query_rec = record.next()
assert query_rec.name == 'query', 'This should be the query'
qstart = int(query_rec.annotations['al_start'])
qstop = int(query_rec.annotations['al_stop'])
gaps_query = 0
qst = query_rec.seq.tostring()
qls = list(qst)
for c in qst.strip('-'):
if c == '-':
gaps_query = gaps_query + 1
# These are for the matching sequences, i.e. the reads
match_rec = record.next()
assert match_rec.name == 'match', 'This should be the match'
mst = match_rec.seq.tostring()
mls = list(mst)
for c in mls:
if c != '-':
mstart = mls.index(c) + 1
break
mstop = len(mst.rstrip('-'))
# counts the gaps in the read (no flanking gaps)
gaps_match = 0
for c in mst.strip('-'):
if c == '-':
gaps_match = gaps_match + 1
match_length = len(mst.strip('-'))
if gaps_query + gaps_match > round(tolerance * match_length):
# print 'too many indels,', (gaps_query + gaps_match)
continue
out_reads[match_rec.id] = [None, None, None, None, []]
out_reads[
match_rec.id][0] = qstart # is this really useful at this time?
out_reads[
match_rec.id][1] = qstop # is this really useful at this time?
out_reads[match_rec.id][2] = mstart # this is
out_reads[match_rec.id][3] = mstop # this too
for i in range(mstart, mstop + 1):
try:
cov_prof[i] = cov_prof[i] + 1
except IndexError:
print 'out of coverage', i
this_q = qls[mstart - 1:mstop]
this_m = list(mst.strip('-'))
assert len(this_q) == len(this_m), 'Length must be the same %d %d' % (
len(this_q), len(this_m))
amb_calls = 0
# There are three possibilities: insertions, deletions, no in-dels
for i in range(len(this_m)):
if this_m[i] == '-' and this_q[i] != '-':
out_reads[match_rec.id][4].append('-')
if this_m[i] != '-' and this_q[i] == '-':
pass
if this_m[i] != '-' and this_q[i] != '-':
out_reads[match_rec.id][4].append(this_m[i])
# This should never happen
if this_m[i] == '-' and this_q[i] == '-':
print 'Should this happen?'
sys.exit()
if this_m[i] == 'N':
amb_calls = amb_calls + 1
if verbose:
print >> sys.stderr, 'Found an N in', match_rec.id
if amb_calls > amb_thresh:
if verbose:
print 'Read', match_rec.id, 'has too many Ns'
del out_reads[match_rec.id]
cp = open('./%s.covprof' % reads_file.rstrip('_reads.fas'), 'w')
for i in range(1, gen_length):
cp.write('%i\t%i\n' % (i, cov_prof[i]))
cp.close()
return out_reads