def alignIndels( all_alleles, colcounts, extend_by = 0 ):
'''align all indel-regions.'''
aa = alignlib.makeAlignatorDPFull( alignlib.ALIGNMENT_LOCAL, 0, 0 )
alignator = alignlib.makeMultipleAlignatorSimple( aa)
ids = all_alleles.keys()
for x,c in enumerate(colcounts):
if c <= 1: continue
sequences = alignlib.StringVector()
for sid in ids:
for allele in all_alleles[sid]:
sequences.append( allele[x] )
mali = alignlib.makeMultAlignment()
alignator.align( mali, sequences )
realigned = []
for line in str(alignlib.MultAlignmentFormatPlain( mali, sequences )).split("\n")[:-1]:
data = line[:-1].split("\t")
realigned.append( data[1] )
assert len(realigned) == len(sequences)
l = max( [len(r) for r in realigned] )
i = 0
for sid in ids:
for allele in all_alleles[sid]:
if realigned[i]: allele[x] = realigned[i]
else: allele[x] = "-" * l
i += 1
colcounts[x] = l
def alignIndels(all_alleles, colcounts, extend_by=0):
'''align all indel-regions.'''
aa = alignlib.makeAlignatorDPFull(alignlib.ALIGNMENT_LOCAL, 0, 0)
alignator = alignlib.makeMultipleAlignatorSimple(aa)
ids = all_alleles.keys()
for x, c in enumerate(colcounts):
if c <= 1: continue
sequences = alignlib.StringVector()
for sid in ids:
for allele in all_alleles[sid]:
sequences.append(allele[x])
mali = alignlib.makeMultAlignment()
alignator.align(mali, sequences)
realigned = []
for line in str(alignlib.MultAlignmentFormatPlain(
mali, sequences)).split("\n")[:-1]:
data = line[:-1].split("\t")
realigned.append(data[1])
assert len(realigned) == len(sequences)
l = max([len(r) for r in realigned])
i = 0
for sid in ids:
for allele in all_alleles[sid]:
if realigned[i]: allele[x] = realigned[i]
else: allele[x] = "-" * l
i += 1
colcounts[x] = l
def _alignToProfile(infile, outfile, min_score=0):
'''align sequences in *infile* against mali
Only alignments with a score higher than *min_score* are accepted.
Output multiple alignment in fasta format to *outfile* and a table
in :file:`outfile.log`.
'''
mali = Mali.Mali()
mali.readFromFile(open("../data/mouse.fasta"))
src_mali = Mali.convertMali2Alignlib(mali)
E.debug("read mali: %i sequences x %i columns" %
(mali.getNumSequences(), mali.getNumColumns()))
# add pseudocounts
profile_mali = mali.getClone()
n = profile_mali.getNumColumns()
for x in "ACGT":
for y in range(0, 2):
profile_mali.addSequence("%s%i" % (x, y), 0, n, x * n)
profile_mali = Mali.convertMali2Alignlib(profile_mali)
alignlib.setDefaultEncoder(alignlib.getEncoder(alignlib.DNA4))
alignlib.setDefaultLogOddor(alignlib.makeLogOddorUniform())
# bg = alignlib.FrequencyVector()
# bg.extend( ( 0.3, 0.1, 0.2, 0.2, 0.2) )
# alignlib.setDefaultRegularizor( alignlib.makeRegularizorTatusov(
# alignlib.makeSubstitutionMatrixDNA4(),
# bg,
# "ACGTN",
# 10.0, 1.0) )
profile = alignlib.makeProfile(profile_mali)
alignment_mode = alignlib.ALIGNMENT_WRAP
alignator = alignlib.makeAlignatorDPFull(alignment_mode, -5.0, -0.5)
map_seq2profile = alignlib.makeAlignmentVector()
map_rseq2profile = alignlib.makeAlignmentVector()
profile.prepare()
# print profile
build_mali = alignlib.makeMultAlignment()
m = alignlib.makeAlignmentVector()
m.addDiagonal(0, n, 0)
build_mali.add(src_mali, m)
outf = open(outfile, "w")
outf_log = open(outfile + ".info", "w")
outf_log.write(
"read_id\tlength\tstart\tend\tparts\tcovered\tpcovered\tscore\tmali_start\tmali_end\tmali_covered\tmali_pcovered\n"
)
sequences, aa = alignlib.StringVector(), alignlib.AlignandumVector()
ids = []
for pid in mali.getIdentifiers():
sequences.append(re.sub("-", "", mali[pid]))
ids.append(pid)
# print str(alignlib.MultAlignmentFormatPlain( build_mali, sequences ))
c = E.Counter()
for s in FastaIterator.FastaIterator(open(infile)):
E.debug("adding %s" % s.title)
c.input += 1
rsequence = Genomics.complement(s.sequence)
seq = alignlib.makeSequence(s.sequence)
rseq = alignlib.makeSequence(rsequence)
alignator.align(map_seq2profile, seq, profile)
alignator.align(map_rseq2profile, rseq, profile)
if map_seq2profile.getScore() > map_rseq2profile.getScore():
m, seq, sequence = map_seq2profile, seq, s.sequence
else:
m, seq, sequence = map_rseq2profile, rseq, rsequence
if m.getLength() == 0:
c.skipped += 1
continue
if m.getScore() < min_score:
c.skipped += 1
continue
r = getParts(m)
covered = 0
for mm in r:
build_mali.add(mm)
sequences.append(sequence)
ids.append(s.title)
covered += mm.getLength() - mm.getNumGaps()
mali_covered = m.getColTo() - m.getColFrom()
outf_log.write("\t".join(
map(str, (s.title, len(s.sequence), m.getRowFrom(), m.getRowTo(),
len(r), covered, "%5.2f" %
(100.0 * covered / len(s.sequence)), m.getScore(),
m.getColFrom(), m.getColTo(), mali_covered, "%5.2f" %
((100.0 * mali_covered) / mali.getNumColumns())))) +
"\n")
c.output += 1
#build_mali.expand( aa )
result = str(
alignlib.MultAlignmentFormatPlain(build_mali, sequences,
alignlib.UnalignedStacked))
for pid, data in zip(ids, result.split("\n")):
start, sequence, end = data.split("\t")
outf.write(">%s/%i-%i\n%s\n" %
(pid, int(start) + 1, int(end), sequence))
outf.close()
outf_log.close()
E.info("%s\n" % str(c))
def _alignToProfile( infile, outfile,
min_score = 0 ):
'''align sequences in *infile* against mali
Only alignments with a score higher than *min_score* are accepted.
Output multiple alignment in fasta format to *outfile* and a table
in :file:`outfile.log`.
'''
mali = Mali.Mali()
mali.readFromFile( open("../data/mouse.fasta") )
src_mali = Mali.convertMali2Alignlib( mali )
E.debug( "read mali: %i sequences x %i columns" % (mali.getNumSequences(), mali.getNumColumns() ))
# add pseudocounts
profile_mali = mali.getClone()
n = profile_mali.getNumColumns()
for x in "ACGT":
for y in range(0,2):
profile_mali.addSequence( "%s%i" % (x,y), 0, n, x * n )
profile_mali = Mali.convertMali2Alignlib( profile_mali )
alignlib.setDefaultEncoder( alignlib.getEncoder( alignlib.DNA4 ) )
alignlib.setDefaultLogOddor( alignlib.makeLogOddorUniform() )
# bg = alignlib.FrequencyVector()
# bg.extend( ( 0.3, 0.1, 0.2, 0.2, 0.2) )
# alignlib.setDefaultRegularizor( alignlib.makeRegularizorTatusov(
# alignlib.makeSubstitutionMatrixDNA4(),
# bg,
# "ACGTN",
# 10.0, 1.0) )
profile = alignlib.makeProfile( profile_mali )
alignment_mode = alignlib.ALIGNMENT_WRAP
alignator = alignlib.makeAlignatorDPFull( alignment_mode,
-5.0,
-0.5 )
map_seq2profile = alignlib.makeAlignmentVector()
map_rseq2profile = alignlib.makeAlignmentVector()
profile.prepare()
# print profile
build_mali = alignlib.makeMultAlignment()
m = alignlib.makeAlignmentVector()
m.addDiagonal( 0, n, 0 )
build_mali.add( src_mali, m )
outf = open( outfile, "w" )
outf_log = open( outfile + ".info", "w" )
outf_log.write( "read_id\tlength\tstart\tend\tparts\tcovered\tpcovered\tscore\tmali_start\tmali_end\tmali_covered\tmali_pcovered\n" )
sequences, aa = alignlib.StringVector(), alignlib.AlignandumVector()
ids = []
for pid in mali.getIdentifiers():
sequences.append( re.sub( "-", "", mali[pid] ) )
ids.append( pid )
# print str(alignlib.MultAlignmentFormatPlain( build_mali, sequences ))
c = E.Counter()
for s in FastaIterator.FastaIterator( open(infile)):
E.debug("adding %s" % s.title )
c.input += 1
rsequence = Genomics.complement(s.sequence)
seq = alignlib.makeSequence( s.sequence )
rseq = alignlib.makeSequence( rsequence )
alignator.align( map_seq2profile, seq, profile )
alignator.align( map_rseq2profile, rseq, profile )
if map_seq2profile.getScore() > map_rseq2profile.getScore():
m, seq, sequence = map_seq2profile, seq, s.sequence
else:
m, seq, sequence = map_rseq2profile, rseq, rsequence
if m.getLength() == 0:
c.skipped += 1
continue
if m.getScore() < min_score:
c.skipped += 1
continue
r = getParts( m )
covered = 0
for mm in r:
build_mali.add( mm )
sequences.append( sequence )
ids.append( s.title )
covered += mm.getLength() - mm.getNumGaps()
mali_covered = m.getColTo() - m.getColFrom()
outf_log.write( "\t".join( map(str, (
s.title,
len(s.sequence),
m.getRowFrom(),
m.getRowTo(),
len(r),
covered,
"%5.2f" % (100.0 * covered / len(s.sequence) ),
m.getScore(),
m.getColFrom(),
m.getColTo(),
mali_covered,
"%5.2f" % ((100.0 * mali_covered) / mali.getNumColumns())
) ) ) + "\n" )
c.output += 1
#build_mali.expand( aa )
result = str(alignlib.MultAlignmentFormatPlain( build_mali,
sequences,
alignlib.UnalignedStacked ))
for pid, data in zip(ids, result.split("\n") ):
start, sequence, end = data.split("\t")
outf.write(">%s/%i-%i\n%s\n" % (pid, int(start)+1, int(end), sequence) )
outf.close()
outf_log.close()
E.info( "%s\n" % str(c) )
