if link.mQueryToken == link.mSbjctToken:
continue
if link.mQueryToken in cds and \
link.mSbjctToken in cds:
# expand to codons
if param_expand:
link.mQueryFrom = (link.mQueryFrom - 1) * 3 + 1
link.mSbjctFrom = (link.mSbjctFrom - 1) * 3 + 1
link.mQueryAli = ScaleAlignment(link.mQueryAli, 3)
link.mSbjctAli = ScaleAlignment(link.mSbjctAli, 3)
map_row2col.clear()
alignlib_lite.AlignmentFormatExplicit(
link.mQueryFrom, link.mQueryAli, link.mSbjctFrom,
link.mSbjctAli).copy(map_row2col)
# test all combinations, the alignment might be a suboptimal alignment in case
# of repeats.
for e1 in cds[link.mQueryToken]:
for e2 in cds[link.mSbjctToken]:
tmp_map_row2col.clear()
if param_expand:
alignlib_lite.copyAlignment(
tmp_map_row2col,
map_row2col,
e1.mPeptideFrom + 1,
e1.mPeptideTo,
e2.mPeptideFrom + 1,
e2.mPeptideTo,
def main(argv=None):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if argv == None: argv = sys.argv
parser = E.OptionParser(
version=
"%prog version: $Id: gpipe/compare_predictions2exons.py 2011 2008-07-04 10:40:51Z andreas $",
usage=globals()["__doc__"])
parser.add_option("-g",
"--genome-file",
dest="genome_file",
type="string",
help="filename with genome.")
parser.add_option("-b",
"--boundaries",
dest="filename_boundaries",
type="string",
help="filename with exon boundaries.")
parser.add_option("-e",
"--exons",
dest="filename_exons",
type="string",
help="filename with exons (output).")
parser.add_option("-p",
"--peptides",
dest="filename_peptides",
type="string",
help="filename with peptide sequences.")
parser.add_option(
"-w",
"--write-notfound",
dest="write_notfound",
action="store_true",
help="print exons for predictions not found in reference.")
parser.add_option("-q",
"--quality-pide",
dest="quality_threshold_pide",
type="int",
help="quality threshold (pide) for exons.")
parser.set_defaults(
genome_file="genome",
filename_boundaries=None,
filename_exons=None,
filename_peptides=None,
quality_threshold_pide=0,
write_notfound=False,
## allowed number of nucleotides for exon boundaries to
## be considered equivalent.
slipping_exon_boundary=9,
## stop codons to search for
stop_codons=("TAG", "TAA", "TGA"),
)
(options, args) = E.Start(parser, add_pipe_options=True)
if len(args) > 0:
print USAGE, "no arguments required."
sys.exit(2)
reference_exon_boundaries = {}
if options.filename_boundaries:
reference_exon_boundaries = Exons.ReadExonBoundaries(open(
options.filename_boundaries, "r"),
do_invert=1,
remove_utr=1)
E.info("read exon boundaries for %i queries" %
len(reference_exon_boundaries))
if options.filename_exons:
outfile_exons = open(options.filename_exons, "w")
outfile_exons.write("%s\n" % "\t".join(
("prediction_id", "exon_id", "exon_from", "exon_to", "exon_frame",
"reference_id", "reference_from", "reference_to",
"reference_phase", "pidentity", "psimilarity", "nframeshifts",
"ngaps", "nstopcodons", "is_ok", "genome_exon_from",
"genome_exon_to")))
else:
outfile_exons = None
if options.filename_peptides:
peptide_sequences = Genomics.ReadPeptideSequences(
open(options.filename_peptides, "r"))
E.info("read peptide sequences for %i queries" %
len(peptide_sequences))
else:
peptide_sequences = {}
entry = PredictionParser.PredictionParserEntry()
last_filename_genome = None
nfound, nmissed_exons, nmissed_length = 0, 0, 0
nempty_alignments = 0
fasta = IndexedFasta.IndexedFasta(options.genome_file)
options.stdout.write("%s\n" % "\t".join(
("prediction_id", "number", "dubious_exons", "boundaries_sum",
"boundaries_max", "identical_exons", "inserted_exons",
"deleted_exons", "inserted_introns", "deleted_introns",
"truncated_Nterminus", "truncated_Cterminus", "deleted_Nexons",
"deleted_Cexons", "inserted_Nexons", "inserted_Cexons")))
for line in sys.stdin:
if line[0] == "#": continue
try:
entry.Read(line)
except ValueError, msg:
print "# parsing failed with msg %s in line %s" % (msg, line[:-1])
sys.exit(1)
exons = Genomics.Alignment2ExonBoundaries(
entry.mMapPeptide2Genome,
query_from=entry.mQueryFrom,
sbjct_from=entry.mSbjctGenomeFrom,
add_stop_codon=0)
if exons[-1][4] != entry.mSbjctGenomeTo:
print "# WARNING: discrepancy in exon calculation!!!"
for e in exons:
print "#", str(e)
print "#", str(entry)
if options.loglevel >= 5:
for e in exons:
print "#", str(e)
genomic_fragment = fasta.getSequence(entry.mSbjctToken,
entry.mSbjctStrand,
entry.mSbjctGenomeFrom,
entry.mSbjctGenomeTo)
skip = False
if peptide_sequences.has_key(entry.mQueryToken):
query_sequence = alignlib_lite.makeSequence(
peptide_sequences[entry.mQueryToken])
sbjct_sequence = alignlib_lite.makeSequence(entry.mTranslation)
percent_similarity, percent_identity = 0, 0
if query_sequence.getLength(
) < entry.mMapPeptide2Translation.getRowTo():
print "# WARNING: query sequence %s is too short: %i %i" % (
entry.mQueryToken, query_sequence.getLength(),
entry.mMapPeptide2Translation.getRowTo())
sys.stdout.flush()
nmissed_length += 1
skip = True
elif sbjct_sequence.getLength(
) < entry.mMapPeptide2Translation.getColTo():
print "# WARNING: sbjct sequence %s is too short: %i %i" % (
entry.mSbjctToken, sbjct_sequence.getLength(),
entry.mMapPeptide2Translation.getColTo())
sys.stdout.flush()
nmissed_length += 1
skip = True
else:
alignlib_lite.rescoreAlignment(
entry.mMapPeptide2Translation, query_sequence,
sbjct_sequence,
alignlib_lite.makeScorer(query_sequence, sbjct_sequence))
percent_identity = alignlib_lite.calculatePercentIdentity(
entry.mMapPeptide2Translation, query_sequence,
sbjct_sequence) * 100
percent_similarity = alignlib_lite.calculatePercentSimilarity(
entry.mMapPeptide2Translation) * 100
E.debug(
"prediction %s: percent identity/similarity: before=%5.2f/%5.2f, realigned=%5.2f/%5.2f"
%
(str(entry.mPredictionId), entry.mPercentSimilarity,
entry.mPercentIdentity, percent_similarity, percent_identity))
else:
query_sequence = None
sbjct_sequence = None
# default values
exons_num_exons = "na"
exons_boundaries_sum = "na"
exons_boundaries_max = "na"
dubious_exons = "na"
ndeleted_exons, ninserted_exons, ndeleted_introns, ninserted_introns, nidentical_exons = 0, 0, 0, 0, 0
truncated_Nterminal_exon, truncated_Cterminal_exon = 0, 0
ndeleted_Nexons, ndeleted_Cexons = 0, 0
ninserted_Nexons, ninserted_Cexons = 0, 0
exons_offset = exons[0][3]
if not reference_exon_boundaries.has_key(entry.mQueryToken):
print "# WARNING: sequence %s has no exon boundaries" % (
entry.mQueryToken)
sys.stdout.flush()
nmissed_exons += 1
skip = True
if not skip:
nfound += 1
ref_exons = reference_exon_boundaries[entry.mQueryToken]
ref_exons_offset = ref_exons[0].mGenomeFrom
exons_num_exons = len(ref_exons) - len(exons)
exons_boundaries_sum = 0
exons_phase = 0
exons_boundaries_max = 0
dubious_exons = 0
inserted_exons = 0
temp_inserted_exons = 0
if options.loglevel >= 3:
for e in exons:
options.stdlog.write("# %s\n" % str(e))
for e in ref_exons:
options.stdlog.write("# %s\n" % str(e))
min_pide = entry.mPercentIdentity * options.quality_threshold_pide / 100
in_sync = 0
e, r = 0, 0
while e < len(exons) and r < len(ref_exons):
this_e, this_r = e + 1, r + 1
percent_identity = 0
percent_similarity = 0
is_good_exon = 0
if options.loglevel >= 4:
options.stdlog.write("# current exons: %i and %i\n" %
(e, r))
sys.stdout.flush()
exon_from, exon_to, exon_phase, exon_genome_from, exon_genome_to, exon_ali = exons[
e][0:6]
ref_from, ref_to, ref_phase, ref_genome_from, ref_genome_to = (
ref_exons[r].mPeptideFrom, ref_exons[r].mPeptideTo,
ref_exons[r].frame, ref_exons[r].mGenomeFrom,
ref_exons[r].mGenomeTo)
ref_genome_from -= ref_exons_offset
ref_genome_to -= ref_exons_offset
## get percent identity for exon
exon_percent_identity = 0
exon_percent_similarity = 0
if query_sequence and sbjct_sequence:
tmp_ali = alignlib_lite.makeAlignmentVector()
xquery_from = exon_from / 3
xquery_to = exon_to / 3
alignlib_lite.copyAlignment(tmp_ali,
entry.mMapPeptide2Translation,
xquery_from, xquery_to)
if tmp_ali.getLength() == 0:
options.stdlog.write(
"# WARNING: empty alignment %s\n" % str(
(ref_from, exon_from, ref_to, exon_to,
xquery_from, xquery_to)))
nempty_alignments += 1
else:
if options.loglevel >= 5:
options.stdlog.write("# %s\n" % str(
alignlib_lite.AlignmentFormatExplicit(
tmp_ali, query_sequence, sbjct_sequence)))
exon_percent_identity = alignlib_lite.calculatePercentIdentity(
tmp_ali, query_sequence, sbjct_sequence) * 100
exon_percent_similarity = alignlib_lite.calculatePercentSimilarity(
tmp_ali) * 100
if exon_percent_identity >= min_pide:
is_good_exon = 1
else:
is_good_exon = 0
if e < len(exons) - 1:
(next_exon_from, next_exon_to, next_exon_phase,
next_exon_genome_from, next_exon_genome_to,
next_exon_ali) = exons[e + 1][0:6]
else:
(next_exon_from, next_exon_to, next_exon_phase,
next_exon_genome_from, next_exon_genome_to,
next_exon_ali) = 0, 0, 0, 0, 0, []
if r < len(ref_exons) - 1:
next_ref_from, next_ref_to, next_ref_phase = (
ref_exons[r + 1].mPeptideFrom,
ref_exons[r + 1].mPeptideTo, ref_exons[r + 1].frame)
else:
next_ref_from, next_ref_to, next_ref_phase = 0, 0, 0
if options.loglevel >= 2:
options.stdlog.write("# %s\n" % "\t".join(
map(str, (entry.mQueryToken, exon_from, exon_to,
exon_phase, exon_genome_from, exon_genome_to,
ref_from, ref_to, ref_phase))))
sys.stdout.flush()
# beware of small exons.
# if less than options.slipping_exon_boundary: boundary is 0
# check if end is more than options.splipping_exon_boundary apart as well.
if exon_to - exon_from <= options.slipping_exon_boundary or \
ref_to - ref_from <= options.slipping_exon_boundary:
boundary = 0
else:
boundary = options.slipping_exon_boundary
if ref_to <= exon_from + boundary and \
ref_to <= exon_to - options.slipping_exon_boundary:
## no overlap
is_good_exon = 0
if e == 0:
ndeleted_Nexons += 1
else:
ndeleted_exons += 1
r += 1
exon_from, exon_to, exon_phase, exon_genome_from, exon_genome_to = 0, 0, 0, 0, 0
overlap = 0
elif exon_to <= ref_from + boundary and \
exon_to <= ref_to - options.slipping_exon_boundary:
## no overlap
is_good_exon = 0
if r == 0:
ninserted_Nexons += 1
else:
ninserted_exons += 1
e += 1
ref_from, ref_to, ref_phase = 0, 0, 0
overlap = 0
else:
## overlap
overlap = 1
dfrom = int(math.fabs(exon_from - ref_from))
dto = int(math.fabs(exon_to - ref_to))
## get percent identity for overlapping fragment
if query_sequence and sbjct_sequence:
## this the problem
tmp_ali = alignlib_lite.makeAlignmentVector()
xquery_from = max(ref_from / 3, exon_from / 3)
xquery_to = min(ref_to / 3, exon_to / 3)
alignlib_lite.copyAlignment(
tmp_ali, entry.mMapPeptide2Translation,
xquery_from, xquery_to)
if tmp_ali.getLength() == 0:
options.stdlog.write(
"# warning: empty alignment %s\n" % str(
(ref_from, exon_from, ref_to, exon_to,
xquery_from, xquery_to)))
percent_identity = 0
percent_similarity = 0
else:
if options.loglevel >= 5:
print str(
alignlib_lite.AlignmentFormatExplicit(
tmp_ali, query_sequence,
sbjct_sequence))
percent_identity = alignlib_lite.calculatePercentIdentity(
tmp_ali, query_sequence, sbjct_sequence) * 100
percent_similarity = alignlib_lite.calculatePercentSimilarity(
tmp_ali) * 100
if percent_identity >= min_pide:
is_good_exon = 1
else:
is_good_exon = 0
dubious_exons += 1
## adjust regions for terminal exons
if e == 0 and r == 0 and dfrom <= (entry.mQueryFrom -
1) * 3 and dfrom > 0:
if is_good_exon:
truncated_Nterminal_exon = dfrom
dfrom = 0
## truncated terminal exons
if e == len(exons) - 1 and r == len(
ref_exons) - 1 and dto <= (
entry.mQueryLength -
entry.mQueryTo) * 3 and dto > 0:
if is_good_exon:
truncated_Cterminal_exon = dto
dto = 0
## do not count deviations for terminal query exons
if e == 0 and dfrom <= entry.mQueryFrom * 3 and dfrom > 0:
dfrom = 0
if e == len(exons) - 1 and dto <= (
entry.mQueryLength -
entry.mQueryTo) * 3 and dto > 0:
dto = 0
## permit difference of one codon (assumed to be stop)
if e == len(exons) - 1 and r == len(
ref_exons) - 1 and dto == 3:
dto = 0
## deal with different boundary conditions:
if dfrom == 0 and dto == 0:
if is_good_exon: nidentical_exons += 1
e += 1
r += 1
## next exon within this ref_exon
elif exon_to < ref_to and next_exon_to and next_exon_to <= ref_to + options.slipping_exon_boundary:
if is_good_exon: ninserted_introns += 1
e += 1
in_sync = 1
dto = 0
## next ref_exon within this exon
elif ref_to < exon_to and next_ref_to and next_ref_to <= exon_to + options.slipping_exon_boundary:
if is_good_exon: ndeleted_introns += 1
r += 1
in_sync = 1
dto = 0
else:
e += 1
r += 1
if in_sync:
dfrom = 0
if is_good_exon:
exons_boundaries_sum += dfrom + dto
exons_boundaries_max = max(dfrom, exons_boundaries_max)
exons_boundaries_max = max(dto, exons_boundaries_max)
###########################################################
## count inserted/deleted introns and misplaced boundaries
##
## if exon and next_exon in ref_exon: inserted intron
## if ref_exon and next_ref_exon in exon: deleted intron
if outfile_exons:
if genomic_fragment and exon_genome_to:
nintrons, nframeshifts, ngaps, nsplits, nstopcodons, disruptions = Genomics.CountGeneFeatures(
exon_genome_from - entry.mSbjctGenomeFrom,
exon_ali,
genomic_fragment,
border_stop_codon=0)
else:
nintrons, nframeshifts, ngaps, nsplits, nstopcodons = 0, 0, 0, 0, 0
if exon_to == 0: this_e = 0
if ref_to == 0: this_r = 0
outfile_exons.write(
string.join(
map(str, (
entry.mPredictionId,
this_e,
exon_from,
exon_to,
exon_phase,
this_r,
ref_from,
ref_to,
ref_phase,
percent_identity,
percent_similarity,
nframeshifts,
ngaps,
nstopcodons,
is_good_exon,
exon_genome_from,
exon_genome_to,
)), "\t") + "\n")
while e < len(exons):
exon_from, exon_to, exon_phase, exon_genome_from, exon_genome_to = exons[
e][0:5]
e += 1
ninserted_Cexons += 1
if outfile_exons:
outfile_exons.write(
string.join(
map(str, (
entry.mPredictionId,
e,
exon_from,
exon_to,
exon_phase,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
exon_genome_from,
exon_genome_to,
)), "\t") + "\n")
while r < len(ref_exons):
ref_from, ref_to, ref_phase, ref_genome_from, ref_genome_to = (
ref_exons[r].mPeptideFrom, ref_exons[r].mPeptideTo,
ref_exons[r].frame, ref_exons[r].mGenomeFrom,
ref_exons[r].mGenomeTo)
ndeleted_Cexons += 1
ref_genome_from -= ref_exons_offset
ref_genome_to -= ref_exons_offset
r += 1
if outfile_exons:
outfile_exons.write(
string.join(
map(str, (
entry.mPredictionId,
0,
0,
0,
0,
r,
ref_from,
ref_to,
ref_phase,
0,
0,
0,
0,
0,
0,
0,
0,
)), "\t") + "\n")
else:
if options.write_notfound:
this_e = 0
## use prediction's identity/similarity for exons.
## This will still then flag stop-codons in later analysis
percent_identity = entry.mPercentIdentity
percent_similarity = entry.mPercentSimilarity
for exon in exons:
this_e += 1
exon_from, exon_to, exon_phase, exon_genome_from, exon_genome_to, exon_ali = exon[
0:6]
if genomic_fragment:
nintrons, nframeshifts, ngaps, nsplits, nstopcodons, disruptions = Genomics.CountGeneFeatures(
exon_genome_from - entry.mSbjctGenomeFrom,
exon_ali, genomic_fragment)
outfile_exons.write(
string.join(
map(str, (
entry.mPredictionId,
this_e,
exon_from,
exon_to,
exon_phase,
0,
0,
0,
0,
percent_identity,
percent_similarity,
nframeshifts,
ngaps,
nstopcodons,
1,
exon_genome_from,
exon_genome_to,
)), "\t") + "\n")
options.stdout.write("\t".join(
map(str, (entry.mPredictionId, exons_num_exons, dubious_exons,
exons_boundaries_sum, exons_boundaries_max,
nidentical_exons, ninserted_exons, ndeleted_exons,
ninserted_introns, ndeleted_introns,
truncated_Nterminal_exon, truncated_Cterminal_exon,
ndeleted_Nexons, ndeleted_Cexons, ninserted_Nexons,
ninserted_Cexons))) + "\n")
def WriteExons(token1, peptide1, cds1, transcript1, token2, peptide2, cds2,
transcript2, peptide_map_a2b):
if param_loglevel >= 3:
for cd in cds1:
print "#", str(cd)
for cd in cds2:
print "#", str(cd)
print "# peptide_map_a2b", str(
alignlib_lite.AlignmentFormatExplicit(peptide_map_a2b))
sys.stdout.flush()
dna_map_a2b = Genomics.AlignmentProtein2CDNA(peptide_map_a2b, cds1, cds2)
if len(cds1) != len(cds2):
if param_loglevel >= 4:
print "" # WARNING: different number of exons!"
seq1 = alignlib_lite.makeSequence(transcript1)
seq2 = alignlib_lite.makeSequence(transcript2)
tmp_map_a2b = alignlib_lite.makeAlignmentVector()
dialign = WrapperDialign.Dialign("-n")
dialignlgs = WrapperDialign.Dialign("-n -it -thr 2 -lmax 30 -smin 8")
dba = WrapperDBA.DBA()
#clustal = WrapperClustal.Clustal()
matrix, gop, gep = global_substitution_matrix
alignator_nw = alignlib_lite.makeAlignatorDPFullDP(
alignlib_lite.ALIGNMENT_GLOBAL, gop, gep, matrix)
alignator_sw = alignlib_lite.makeAlignatorDPFullDP(
alignlib_lite.ALIGNMENT_LOCAL, gop, gep, matrix)
# concatenated alignments for exons:
# 1: only the common parts
ali_common1 = ""
ali_common2 = ""
e1, e2 = 0, 0
while cds1[e1].mGenomeTo <= dna_map_a2b.getRowFrom():
e1 += 1
while cds2[e2].mGenomeTo <= dna_map_a2b.getColFrom():
e2 += 1
nskipped, nerrors = 0, 0
if param_loglevel >= 5:
nmapped = 0
for x in range(dna_map_a2b.getRowFrom(), dna_map_a2b.getRowTo() + 1):
if dna_map_a2b.mapRowToCol(x) >= 0:
nmapped += 1
print "# nmapped=", nmapped
print str(alignlib_lite.AlignmentFormatEmissions(dna_map_a2b))
# declare alignments used
map_intron_a2b = alignlib_lite.makeAlignmentVector()
result = Exons.CompareGeneStructures(cds1,
cds2,
map_cmp2ref=peptide_map_a2b)
if param_loglevel >= 2:
print result.Pretty("#")
nskipped_exons, nskipped_introns = 0, 0
last_e1, last_e2 = None, None
for link in result.mEquivalences:
if link.mCoverage <= param_min_exon_coverage:
nskipped_exons += 1
continue
e1, e2 = link.mId1, link.mId2
c1 = cds1[e1]
c2 = cds2[e2]
exon_fragment1 = transcript1[c1.mGenomeFrom:c1.mGenomeTo]
exon_fragment2 = transcript2[c2.mGenomeFrom:c2.mGenomeTo]
#######################################################################
# write unaligned exons
if param_write_exons:
pair = AlignedPairs.UnalignedPair()
pair.mCategory = "exon"
pair.mToken1 = token1
pair.mId1 = e1 + 1
pair.mNum1 = len(cds1)
pair.mLen1 = len(exon_fragment1)
pair.mSequence1 = exon_fragment1
pair.mToken2 = token2
pair.mId2 = e2 + 1
pair.mNum2 = len(cds2)
pair.mLen2 = len(exon_fragment2)
pair.mSequence2 = exon_fragment2
pair.mFrom1, pair.mTo1 = c1.mGenomeFrom, c1.mGenomeTo,
pair.mFrom2, pair.mTo2 = c2.mGenomeFrom, c2.mGenomeTo,
print str(pair)
sys.stdout.flush()
#######################################################################
# build alignment for overlap of both exons
# tmp_map_a2b.clear()
# alignlib_lite.copyAlignment( tmp_map_a2b, dna_map_a2b,
# c1.mGenomeFrom + 1, c1.mGenomeTo )
# if param_loglevel >= 5:
# print "# alignment: %i-%i" % (c1.mGenomeFrom + 1, c1.mGenomeTo)
# for x in alignlib_lite.writeAlignmentTable( tmp_map_a2b ).split("\n"):
# print "#", x
# if tmp_map_a2b.getLength() == 0:
# if param_loglevel >= 1:
# print "# WARNING: empty alignment between exon %i (from %i to %i) and exon %i" % \
## (e1,c1.mGenomeFrom + 1, c1.mGenomeTo, e2)
# print "## peptide_map_a2b", peptide_map_a2b.getRowFrom(), peptide_map_a2b.getRowTo(),\
## peptide_map_a2b.getColFrom(), peptide_map_a2b.getColTo(), \
# Alignlib.writeAlignmentCompressed(peptide_map_a2b)
# print "## dna_map_a2b", dna_map_a2b.getRowFrom(), dna_map_a2b.getRowTo(),\
## dna_map_a2b.getColFrom(), dna_map_a2b.getColTo(), \
# Alignlib.writeAlignmentCompressed(dna_map_a2b)
# for cd in cds1: print "##", str(cd)
# for cd in cds2: print "##", str(cd)
## nerrors += 1
# continue
## data = map(lambda x: x.split("\t"), alignlib_lite.writePairAlignment( seq1, seq2, tmp_map_a2b ).split("\n"))
# if "caligned" in param_write_exons :
# print "exon\tcaligned\t%s\t%i\t%s\t%i\t%s\t%s\t%s\t%s\t%s\t%s" % ( token1, e1,
## token2, e2,
## data[0][0], data[0][2],
## data[1][0], data[1][2],
# data[0][1], data[1][1] )
## ali_common1 += data[0][1]
## ali_common2 += data[1][1]
#######################################################################
# write alignment of introns for orthologous introns
# orthologous introns are between orthologous exons
if param_write_introns:
if last_e1 is not None:
if e1 - last_e1 != 1 or e2 - last_e2 != 1:
nskipped_introns += 1
else:
pair = AlignedPairs.UnalignedPair()
intron_from1 = cds1[e1 - 1].mGenomeTo
intron_to1 = cds1[e1].mGenomeFrom
intron_from2 = cds2[e2 - 1].mGenomeTo
intron_to2 = cds2[e2].mGenomeFrom
intron_fragment1 = transcript1[intron_from1:intron_to1]
intron_fragment2 = transcript2[intron_from2:intron_to2]
if len(intron_fragment1) == 0 or len(
intron_fragment2) == 0:
print "## ERROR: empty intron fragments: %i-%i out of %i and %i-%i out of %i." %\
(intron_from1, intron_to1, len(transcript1),
intron_from2, intron_to2, len(transcript2))
continue
pair.mCategory = "intron"
pair.mToken1 = token1
pair.mId1 = e1 + 1
pair.mNum1 = len(cds1) - 1
pair.mLen1 = len(intron_fragment1)
pair.mFrom1 = intron_from1
pair.mTo1 = intron_to1
pair.mSequence1 = intron_fragment1
pair.mToken2 = token2
pair.mId2 = e2 + 1
pair.mNum1 = len(cds2) - 1
pair.mLen2 = len(intron_fragment2)
pair.mFrom2 = intron_from2
pair.mTo2 = intron_to2
pair.mSequence2 = intron_fragment2
if (param_min_intron_length and len(intron_fragment1) < param_min_intron_length) or \
(param_min_intron_length and len(intron_fragment2) < param_min_intron_length) or \
(param_max_intron_length and len(intron_fragment1) > param_max_intron_length) or \
(param_max_intron_length and len(intron_fragment2) > param_max_intron_length):
if param_loglevel >= 1:
print "# skipped: fragment lengths out of bounds for: %s\t%s\t%s\t%s\t%i\t%i" %\
(token1, e1, token2, e2,
len(intron_fragment1),
len(intron_fragment2))
sys.stdout.flush()
nskipped += 1
print str(pair)
# else:
## anchored_from1 = intron_from1 - param_extend_introns
## anchored_to1 = intron_to1 + param_extend_introns
## anchored_from2 = intron_from2 - param_extend_introns
## anchored_to2 = intron_to2 + param_extend_introns
## anchored_fragment1 = transcript1[anchored_from1:anchored_to1]
## anchored_fragment2 = transcript2[anchored_from2:anchored_to2]
# for method in param_write_introns:
# if param_loglevel >= 2:
# print "## aligning with method %s" % method
# sys.stdout.flush
# map_intron_a2b.clear()
# if method == "unaligned":
## from1, to1, ali1, from2, to2, ali2 = 0, 0, intron_fragment1, 0, 0, intron_fragment2
# elif method in ("dialigned", "dbaligned", "clusaligned", "dialignedlgs"):
## tmp_intron_a2b = alignlib_lite.makeAlignmentVector()
# if param_loglevel >= 1:
# print "# aligning with method %s two fragments of length %i and %i" % (method,
# len(anchored_fragment1),
# len(anchored_fragment2))
# sys.stdout.flush()
# if method == "dialigned":
## result = dialign.Align( anchored_fragment1, anchored_fragment2, tmp_intron_a2b )
# elif method == "dialignedlgs":
## result = dialignlgs.Align( anchored_fragment1, anchored_fragment2, tmp_intron_a2b )
# elif method == "dbaligned":
## result = dba.Align( anchored_fragment1, anchored_fragment2, tmp_intron_a2b )
# elif method == "clusaligned":
## result = clustal.Align( anchored_fragment1, anchored_fragment2, tmp_intron_a2b )
# if not result or result.getLength() == 0:
# if param_loglevel >= 1:
# print "# Error: empty intron alignment"
# sys.stdout.flush()
## nerrors += 1
# continue
## tmp_intron_a2b.moveAlignment( anchored_from1, anchored_from2 )
# alignlib_lite.copyAlignment( map_intron_a2b, tmp_intron_a2b,
## intron_from1 + 1, intron_to1,
# intron_from2 + 1, intron_to2 )
# elif method == "nwaligned":
## seq1.useSegment( cds1[e1-1].mGenomeTo + 1, cds1[e1].mGenomeFrom )
## seq2.useSegment( cds2[e2-1].mGenomeTo + 1, cds2[e2].mGenomeFrom )
## alignator_nw.Align( seq1, seq2, map_intron_a2b )
# seq1.useFullLength()
# seq2.useFullLength()
# elif method == "swaligned":
## seq1.useSegment( cds1[e1-1].mGenomeTo + 1, cds1[e1].mGenomeFrom )
## seq2.useSegment( cds2[e2-1].mGenomeTo + 1, cds2[e2].mGenomeFrom )
## alignlib_lite.performIterativeAlignment( map_intron_a2b, seq1, seq2, alignator_sw, param_min_score_sw )
# seq1.useFullLength()
# seq2.useFullLength()
# else:
## raise "unknown method %s" % method
# if map_intron_a2b.getLength() > 0:
# if param_compress:
## from1, to1 = map_intron_a2b.getRowFrom(), map_intron_a2b.getRowTo()
## from2, to2 = map_intron_a2b.getColFrom(), map_intron_a2b.getColTo()
## ali1, ali2 = Alignlib.writeAlignmentCompressed( map_intron_a2b )
# else:
# data = map(lambda x: x.split("\t"),
# alignlib_lite.writePairAlignment( seq1, seq2, map_intron_a2b ).split("\n"))
# if len(data) < 2:
## data=[ ( 0, "", 0), (0, "", 0)]
## from1, ali1, to1 = data[0]
## from2, ali2, to2 = data[1]
# print string.join(map(str, ("intron",
# method,
## token1, e1, len(cds1) - 1, len(intron_fragment1),
## token2, e2, len(cds2) - 1, len(intron_fragment2),
# map_intron_a2b.getNumGaps(),
# map_intron_a2b.getLength(),
## map_intron_a2b.getLength() - map_intron_a2b.getNumGaps(),
## from1, to1, ali1,
## from2, to2, ali2,
## intron_from1, intron_to1,
# intron_from2, intron_to2)), "\t")
# sys.stdout.flush()
last_e1, last_e2 = e1, e2
##########################################################################
# write concatenated exons
# for method in param_write_exons:
# if method == "common":
# print "exon\tcommon\t%s\t%i\t%s\t%i\t%s\t%s\t%s\t%s\t%s\t%s" % ( token1, 0,
## token2, 0,
## 0, 0,
## 0, 0,
# ali_common1, ali_common2 )
# elif method == "exons":
# Write full alignment without gaps.
# This will not care about exon boundaries and gaps.
# data = map(lambda x: x.split("\t"),
# alignlib_lite.writePairAlignment( seq1, seq2, dna_map_a2b ).split("\n"))
# try:
## from1, s1, to1, from2, s2, to2 = data[0] + data[1]
# except ValueError:
## from1, to1, from2, to2 = 0, 0, 0, 0
## s1, s2 = "", ""
## nerrors += 1
# except IndexError:
## from1, to1, from2, to2 = 0, 0, 0, 0
## s1, s2 = "", ""
## nerrors += 1
# if from1:
# if len(s1) != len(s2):
# print "# WARNING: alignment of different lengths: %i and %i" % (len(s1), len(s2))
## nerrors += 1
## from1, to1, from2, to2 = 0, 0, 0, 0
## s1, s2 = "", ""
# else:
## a1, a2 = [], []
# for x in range( min(len(s1), len(s2)) ):
# if s1[x] != "-" and s2[x] != "-":
## a1.append( s1[x] )
## a2.append( s2[x] )
## s1 = string.join(a1, "")
## s2 = string.join(a2, "")
# print "exon\texons\t%s\t%i\t%s\t%i\t%s\t%s\t%s\t%s\t%s\t%s" % ( (token1, 0,
## token2, 0,
## from1, to1,
## from2, to2,
# s1, s2 ) )
# elif method == "full":
# write full alignment (do not care about exon boundaries)
# data = map(lambda x: x.split("\t"),
# alignlib_lite.writePairAlignment( seq1, seq2, dna_map_a2b ).split("\n"))
## if len(data) < 2: data=[ ( 0, "", 0), (0, "", 0)]
# print "exon\tfull\t%s\t%i\t%s\t%i\t%s\t%s\t%s\t%s\t%s\t%s" % ( token1, 0,
## token2, 0,
## data[0][0], data[0][2],
## data[1][0], data[1][2],
# data[0][1], data[1][1] )
if param_loglevel >= 3:
print "# skipped_exons=%i, skipped_introns=%i" % (nskipped_exons,
nskipped_introns)
return nerrors, nskipped
def main( argv = None ):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if argv == None: argv = sys.argv
parser.add_option("-m", "--filename-map", dest="filename_map", type="string",
help="filename with mapping information.")
parser.add_option("-o", "--pattern-old", dest="pattern_old", type="string",
help="pattern for mapping new to old identifiers: extract string from old.")
parser.add_option("-n", "--pattern-new", dest="pattern_new", type="string",
help="pattern for mapping new to old identifiers: put string into new.")
parser.add_option("-g", "--genome-file", dest="genome_file", type="string",
help="genome_file.")
parser.add_option("-p", "--peptides", dest="filename_peptides", type = "string",
help="filename with peptide sequences.")
parser.add_option("-f", "--input-format", dest="input_format", type="choice",
help="format of mapping file", choices=("alignment", "offsets") )
parser.add_option("-i", "--write-missed", dest="write_missed", type="string",
help="write missed identifiers to separate file.")
parser.add_option("-a", "--filename-genes", dest="filename_genes", type="string",
help="filename with gene information.")
parser.add_option("--filename-old-peptides", dest="filename_old_peptides", type="string",
help="filename with old peptide information.")
parser.add_option("--no-renumber", dest="renumber", action="store_false",
help="do not renumber predictions.")
parser.add_option("--contig-sizes-old", dest="contig_sizes_old", type="string",
help="contig sizes for old data.")
parser.add_option("--contig-sizes-new", dest="contig_sizes_new", type="string",
help="contig sizes for new data.")
parser.add_option("--skip-errors", dest="skip_errors", action="store_true",
help="skip entries with errors.")
parser.set_defaults(
filename_map = None,
pattern_old = "(.+)",
pattern_new = "%s",
genome_file = None,
filename_peptides = None,
write_missed = None,
filename_genes = None,
filename_old_peptides = None,
renumber = True,
input_format = "alignment",
contig_sizes_old = None,
contig_sizes_new = None,
skip_errors = None
)
(options, args) = E.Start( parser, add_pipe_options = True)
predictor = PredictorExonerate()
## the different mapping criteria
map_sbjcts = {}
breakpoints = {}
################################################################################################
map_transcript2gene = {}
if options.filename_genes:
infile = open(options.filename_genes, "r")
for gene, transcript in map( lambda x: x[:-1].split("\t")[:2], filter( lambda x: x[0] != "#", infile.readlines())):
map_transcript2gene[transcript] = gene
infile.close()
################################################################################################
peptides = {}
if options.filename_peptides:
peptides = Genomics.ReadPeptideSequences( open(options.filename_peptides, "r"))
options.stdlog.write( "# read %i peptide sequences.\n" % len(peptides))
################################################################################################
## read old query sequences and compare against new query sequences
## this can be used to build a map between old and new queries
query_map_old2new = {}
if options.filename_old_peptides:
old_peptides = Genomics.ReadPeptideSequences( open(options.filename_old_peptides, "r"))
options.stdlog.write( "# read %i old peptide sequences.\n" % len(old_peptides))
query_map_old2new, unmappable, unmapped = Genomics.MapSequences( old_peptides, peptides)
options.stdlog.write( "# built map: unmappable=%i unmapped=%i.\n" % (len(unmappable), len(unmapped)))
if options.loglevel >= 2:
options.stdlog.write( "# unmappable: %s.\n" % ";".join(unmappable))
options.stdlog.write( "# unmapped: %s.\n" % ";".join(unmapped))
################################################################################################
## read old/new contig sizes for mapping positive/negative coordinates
contig_sizes_old = {}
contig_sizes_new = {}
if options.contig_sizes_old:
contig_sizes_old = Genomics.ReadContigSizes( open(options.contig_sizes_old, "r") )
if options.contig_sizes_new:
contig_sizes_new = Genomics.ReadContigSizes( open(options.contig_sizes_new, "r") )
################################################################################################
if options.filename_map:
infile = open(options.filename_map)
if options.input_format == "alignments":
for line in infile:
if line[0] == "#": continue
x, old_token, old_from, old_to, old_ali, new_from, new_to, new_ali = line[:-1].split("\t")
map_sbjcts[old_token] = (old_from, old_ali, new_from, new_ali)
if options.loglevel >= 1:
options.stdlog.write( "# read %i alignments.\n" % len(map_sbjcts))
elif options.input_format == "offsets":
## input is a list of segments and their offsets.
breakpoints, endpoints, offsets = ReadOffsets( infile )
if options.loglevel >= 1:
options.stdlog.write( "# read breakpoints for %i chromosomes.\n" % len(breakpoints))
infile.close()
################################################################################################
################################################################################################
################################################################################################
## end of input section
################################################################################################
################################################################################################
################################################################################################
rx = re.compile(options.pattern_old)
last_sbjct_token = None
ninput = 0
nerrors = 0
nerrors_map = 0
nerrors_inconsistencies = 0
nerrors_boundaries = 0
nerrors_translation = 0
nerrors_inconsequential = 0
nerrors_realigned = 0
nmapped = 0
nfiltered = 0
naligned = 0
noutput = 0
found_transcripts = {}
nduplicates = 0
output = {}
for line in sys.stdin:
if line[0] == "#": continue
entry = PredictionParser.PredictionParserEntry()
entry.Read( line )
ninput += 1
is_positive = entry.mSbjctStrand == "+"
is_error = False
## check if query token is mappable: using sequence map
if (query_map_old2new and entry.mQueryToken not in query_map_old2new):
options.stdlog.write("# skipping prediction %i: obsolete query %s\n" % (entry.mPredictionId, entry.mQueryToken) )
nfiltered += 1
continue
else:
## check if query token is mappable: using filter
if (peptides and entry.mQueryToken not in peptides):
options.stdlog.write("# skipping prediction %i: obsolete query %s\n" % (entry.mPredictionId, entry.mQueryToken) )
nfiltered += 1
continue
new_sbjct_token = options.pattern_new % rx.search(entry.mSbjctToken).groups()[0]
##########################################################################################################
## Map via alignments
if entry.mSbjctToken in map_sbjcts:
nmapped += 1
if last_sbjct_token != entry.mSbjctToken:
old_from, old_ali, new_from, new_ali = map_sbjcts[entry.mSbjctToken]
map_a2b = alignlib_lite.makeAlignmentVector()
alignlib_lite.AlignmentFormatExplicit(
int(old_from), old_ali,
int(new_from), new_ali).copy( map_a2b )
last_sbjct_token = entry.mSbjctToken
if options.loglevel >= 3:
print "#", str(entry)
print "#", map_sbjcts[entry.mSbjctToken]
sys.stdout.flush()
old_f, old_t = entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo
## convert to forward coordinates:
if is_positive:
f, t= old_f, old_t
first_res, last_res = f + 1, t
else:
f, t = map_a2b.getRowTo() - old_f, map_a2b.getRowTo() - old_t
first_res, last_res = f, t + 1
## map first and last residues
mfirst_res = map_a2b.mapRowToCol( first_res )
mlast_res = map_a2b.mapRowToCol( last_res )
if (mfirst_res == 0 and old_f != 0) or (mlast_res == 0 and old_t != map_a2b.getRowTo() ):
options.stderr.write("# mapping not possible for prediction %i on %s %s:%i-%i -> %i-%i -> %i-%i -> %i-%i -> %i-%i\n" % \
(entry.mPredictionId, entry.mSbjctToken, entry.mSbjctStrand,
old_f, old_t,
f, t,
first_res, last_res,
mfirst_res, mlast_res,
f, t))
options.stderr.write("# %s\n" % str(map_sbjcts[entry.mSbjctToken]))
options.stderr.write("# %s\n" % str(entry))
options.stderr.flush()
nerrors_boundaries += 1
is_error = True
## get extended boundaries for alignment later on
while mfirst_res == 0 and first_res > 1:
first_res -= 1
mfirst_res = map_a2b.mapRowToCol(first_res)
while mlast_res == 0 and last_res < map_a2b.getRowTo():
last_res += 1
mlast_res = map_a2b.mapRowToCol(last_res)
## convert to genomic coordinates
## convert negative strand coordinates
if is_positive:
new_f = mfirst_res - 1
new_t = mlast_res
else:
new_f = mfirst_res
new_t = mlast_res - 1
new_f = map_a2b.getColTo() - new_f
new_t = map_a2b.getColTo() - new_t
## Now map the alignment.
try:
MapAlignment( entry, map_a2b )
except ValueError:
options.stderr.write("# alignment mapping not possible for prediction %i on %s %s:%i-%i -> %i-%i -> %i-%i -> %i-%i -> %i-%i -> %i-%i\n" % \
(entry.mPredictionId, entry.mSbjctToken, entry.mSbjctStrand,
old_f, old_t,
f, t,
first_res, last_res,
mfirst_res, mlast_res,
new_f, new_t,
entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo))
options.stderr.write("# %s\n" % str(map_sbjcts[entry.mSbjctToken]))
options.stderr.flush()
nerrors_map += 1
is_error= True
if new_f != entry.mSbjctGenomeFrom or new_t != entry.mSbjctGenomeTo:
options.stderr.write("# mapping inconsistency for prediction %i on %s %s:%i-%i -> %i-%i -> %i-%i -> %i-%i -> %i-%i <> %i-%i\n" % \
(entry.mPredictionId, entry.mSbjctToken, entry.mSbjctStrand,
old_f, old_t,
f, t,
first_res, last_res,
mfirst_res, mlast_res,
new_f, new_t,
entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo))
nerrors_inconsistencies += 1
is_error = True
##########################################################################################################
## Map via offsets
if entry.mSbjctToken in breakpoints:
old_f, old_t = entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo
## convert to forward coordinates:
if is_positive:
f, t= old_f, old_t
else:
f, t = contig_sizes_old[entry.mSbjctToken] - old_t, contig_sizes_old[entry.mSbjctToken] - old_f
o1 = GetOffset( f,
breakpoints[entry.mSbjctToken],
endpoints[entry.mSbjctToken],
offsets[entry.mSbjctToken] )
o2 = GetOffset( t,
breakpoints[entry.mSbjctToken],
endpoints[entry.mSbjctToken],
offsets[entry.mSbjctToken] )
if o1 != o2:
options.stderr.write("# break within gene %s\n" % str(entry))
nerrors_map += 1
is_error = True
f += o1
t += o2
if not is_positive:
f, t = contig_sizes_new[entry.mSbjctToken] - t, contig_sizes_new[entry.mSbjctToken] - f
entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo = f, t
if entry.mSbjctGenomeFrom > entry.mSbjctGenomeTo:
options.stderr.write("# mapping error: start after end %s\n" % str(entry))
nerrors_map += 1
is_error = True
##########################################################################################################
## do translation check, if genome is given
if options.genome_file:
genomic_sequence = Genomics.GetGenomicSequence( new_sbjct_token, entry.mSbjctStrand,
entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo,
options.genome_file,
loglevel = 0)
map_peptide2translation, translation = Genomics.Alignment2PeptideAlignment( \
entry.mMapPeptide2Genome, entry.mQueryFrom, 0, genomic_sequence )
if re.sub("X", "", translation) != re.sub("X", "", entry.mTranslation):
options.stderr.write("# translation error for prediction %i on %s %s:%i-%i -> %i-%i <> %i-%i\n" % \
(entry.mPredictionId, entry.mSbjctToken, entry.mSbjctStrand,
old_f, old_t,
f, t,
entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo))
if map_sbjcts:
options.stderr.write("# %s\n" % str(map_sbjcts[entry.mSbjctToken]))
options.stderr.write("# old=%s\n# new=%s\n" % (entry.mTranslation, translation))
options.stderr.write("# old=%s\n# new=%s\n" % (entry.mAlignmentString, Genomics.Alignment2String(entry.mMapPeptide2Genome)))
nerrors_translation += 1
is_error = True
if peptides and entry.mQueryToken in peptides:
naligned += 1
options.stdlog.write( "# aligning: %s versus %s:%s: %i-%i\n" % ( \
entry.mQueryToken,
new_sbjct_token, entry.mSbjctStrand,
entry.mSbjctGenomeFrom, entry.mSbjctGenomeTo))
# do a quick reprediction
if entry.mQueryToken in peptides:
genomic_sequence = Genomics.GetGenomicSequence( new_sbjct_token, entry.mSbjctStrand,
0, 0,
genome_file = options.genome_pattern,
loglevel = 0)
predictor.mLogLevel = 0
result = predictor(entry.mQueryToken, peptides[entry.mQueryToken],
entry.mSbjctToken, genomic_sequence,
"--exhaustive --subopt FALSE --score '%s' " % str(80),
new_f - 10, new_t + 10)
prediction_id = entry.mPredictionId
if result:
entry = result[0]
entry.mPredictionId = prediction_id
nerrors_realigned += 1
else:
if is_error:
nerrors_inconsequential += 1
entry.mSbjctToken = new_sbjct_token
## map query tokens
if query_map_old2new:
query_tokens = query_map_old2new[entry.mQueryToken]
else:
query_tokens = (entry.mQueryToken,)
if options.skip_errors and is_error:
continue
for query_token in query_tokens:
entry.mQueryToken = query_token
prediction_id = entry.mPredictionId
entry.mPredictionId = 0
hid = Genomics.GetHID( str(entry) )
if hid in output:
nduplicates += 1
continue
noutput += 1
if options.renumber: prediction_id = noutput
entry.mPredictionId = prediction_id
options.stdout.write( str(entry) + "\n")
options.stdout.flush()
found_transcripts[entry.mQueryToken] = 1
## write out found transcripts and genes
nmissed_transcripts = 0
missed_transcripts = []
found_genes = {}
if peptides:
for x in peptides.keys():
if x not in found_transcripts:
nmissed_transcripts += 1
missed_transcripts.append( x )
else:
found_genes[map_transcript2gene[x]] = 1
missed_genes = {}
nmissed_genes = 0
if map_transcript2gene:
for t in missed_transcripts:
g = map_transcript2gene[t]
if g not in found_genes:
missed_genes[g] = 1
nmissed_genes = len(missed_genes)
if options.write_missed:
outfile = open(options.write_missed, "w")
for x in missed_transcripts:
if x in unmapped:
status = "unmapped"
else:
status = "mapped"
outfile.write( "%s\t%s\t%s\n" % ("transcript", x, status ))
for x in missed_genes:
status = "unknown"
outfile.write( "%s\t%s\t%s\n" % ("gene", x, status ))
outfile.close()
options.stdlog.write("# input=%i, output=%i, filtered=%i, nduplicates=%i, mapped=%i, errors=%i\n" % (\
ninput, noutput, nfiltered, nduplicates, nmapped, nerrors ))
options.stdlog.write("# errors: inconsequental=%i, boundaries=%i, mapping=%i, inconsistencies=%i, translation=%i, realigned=%i\n" % (\
nerrors_inconsequential, nerrors_boundaries, nerrors_map, nerrors_inconsistencies, nerrors_translation, nerrors_realigned ))
options.stdlog.write("# peptides: input=%i, found=%i, missed=%i, found_genes=%i, missed_genes=%i\n" % (\
len(peptides), len(found_transcripts), nmissed_transcripts, len(found_genes), nmissed_genes) )
E.Stop()
def main( argv = None ):
"""script main.
parses command line options in sys.argv, unless *argv* is given.
"""
if argv == None: argv = sys.argv
parser = E.OptionParser( version = "%prog version: $Id: gpipe/prediction2pairs.py 2031 2008-07-15 09:19:05Z andreas $", usage = globals()["__doc__"])
parser.add_option( "-g", "--genome-file", dest="genome_file", type="string",
help="filename with genomic data (indexed)." )
parser.add_option( "-c", "--cds", dest="filename_cds", type="string",
help="filename with cds seguences." )
parser.add_option( "-f", "--format", dest="format", type="choice",
choices=("paired_fasta", ),
help="output format, valid options are: paired_fasta: concatenated pairwise alignments in FASTA format" )
parser.set_defaults(
genome_file = "genome",
filename_cds = "cds.fasta",
format = "paired_fasta",
filename_suffix = ".fasta",
filename_prefix = "",
)
(options, args) = E.Start( parser, add_psql_options = True )
if len(args) > 0:
print USAGE, "no arguments required."
sys.exit(1)
fasta = IndexedFasta.IndexedFasta( options.genome_file )
## reading CDS sequences
if options.filename_cds:
cds_sequences = Genomics.ReadPeptideSequences( open(options.filename_cds, "r") )
else:
cds_sequences = {}
if options.loglevel >= 1:
options.stdlog.write( "# read %i CDS sequences\n" % len(cds_sequences) )
last_filename_genome = None
p = PredictionParser.PredictionParserEntry()
ninput, noutput, nsanity, n3, nlength = 0, 0, 0, 0, 0
for line in options.stdin:
if line[0] == "#": continue
if line[0] == '"': continue
p.Read(line)
ninput += 1
genomic_fragment = fasta.getSequence( p.mSbjctToken, p.mSbjctStrand,
p.mSbjctGenomeFrom, p.mSbjctGenomeTo )
if len(genomic_fragment) == 0:
raise "ERROR: empty fragment %s:%s for line" % (p.mSbjctGenomeFrom, p.mSbjctGenomeTo), line
try:
cds_fragment = cds_sequences[p.mQueryToken]
except KeyError:
options.stdlog.write( "# ERROR: cds not found: query %s.\n" % p.mQueryToken )
continue
map_query2sbjct, genomic_fragment = Genomics.Alignment2CDNA( p.mMapPeptide2Genome,
query_from = p.mQueryFrom,
sbjct_from = 0,
genome = genomic_fragment )
## check for errors:
if map_query2sbjct.getRowTo() != p.mQueryTo * 3:
options.stdlog.write( "# ERROR: boundary shift in query at line %s\n# %i %i\n" % (line, map_query2sbjct.getRowTo(), p.mQueryTo * 3 ) )
if map_query2sbjct.getColTo() > len(genomic_fragment):
options.stdlog.write( "# ERROR: length mismatch in line %s\n# genomic fragment (%i) shorter than last aligned residue (%i)\n" %\
(line, len(genomic_fragment), map_query2sbjct.getColTo()) )
options.stdlog.write( "# cds %s\n# genomic %s\n" % (str( cds_fragment ), genomic_fragment ))
nlength += 1
continue
if map_query2sbjct.getRowTo() > len(cds_fragment):
options.stdlog.write( "# ERROR: length mismatch in line %s\n# cds fragment (%i) shorter than last aligned residue (%i)\n" %\
(line, len(cds_fragment), map_query2sbjct.getRowTo()) )
options.stdlog.write( "# cds %s\n# genomic %s\n" % (str( cds_fragment ), genomic_fragment ))
nlength += 1
continue
cds_seq = alignlib_lite.makeSequence( cds_fragment )
genomic_seq = alignlib_lite.makeSequence( genomic_fragment )
f = alignlib_lite.AlignmentFormatExplicit( map_query2sbjct, cds_seq, genomic_seq )
row_ali = f.mRowAlignment
col_ali = f.mColAlignment
row_ali, col_ali = Genomics.RemoveFrameShiftsFromAlignment(row_ali, col_ali)
row_ali = Genomics.MaskStopCodons( row_ali )
col_ali = Genomics.MaskStopCodons( col_ali )
if len(row_ali) != len(col_ali):
options.stdlog.write( "# ERROR: wrong alignment lengths.\n" )
sys.exit(1)
if len(row_ali) % 3 or len(col_ali) % 3:
options.stdlog.write( "# ERROR: sequences are not a multiple of 3 in line: %s\n" % line )
options.stdlog.write( "# %6i %s\n# %6i %s\n" % (len(row_ali), str(row_ali), len(col_ali), str(col_ali) ) )
n3 += 1
input = re.sub( "[-X]", "", p.mTranslation )
ref = re.sub( "[-X]", "", Genomics.TranslateDNA2Protein( col_ali ) )
if input != ref:
if options.loglevel >= 1:
options.stdlog.write("# sanity check failed for %s - %s\n# %6i %s\n# %6i %s\n" % (p.mPredictionId, p.mQueryToken,
len(input), input,
len(ref), ref ) )
nsanity += 1
continue
options.stdout.write( ">%s\n%s\n" % (p.mPredictionId, row_ali) )
options.stdout.write( ">%s_vs_%s_%s_%i_%i\n%s\n" % \
(p.mQueryToken, p.mSbjctToken, p.mSbjctStrand, p.mSbjctGenomeFrom, p.mSbjctGenomeTo, col_ali) )
noutput += 1
if options.loglevel >= 1:
options.stdlog.write("# ninput=%i, noutput=%i, nsanity=%i, nlength=%i, n3=%i\n" % (ninput, noutput, nsanity, nlength, n3) )
E.Stop()
def EliminateRedundantEntries( rep,
data,
eliminated_predictions,
options,
peptides,
extended_peptides,
filter_quality = None,
this_quality = None ):
"""eliminate redundant entries in a set."""
eliminated = []
rep_id = rep.transcript_id
rep_coverage, rep_pid = rep.mQueryCoverage, rep.mPid
alignator = alignlib_lite.makeAlignatorDPFull( alignlib_lite.ALIGNMENT_LOCAL, options.gop, options.gep )
result = alignlib_lite.makeAlignmentVector()
rep_seq = peptides[rep_id]
rep_extended_seq = extended_peptides[rep_id]
for entry in data:
mem_id, mem_coverage, mem_pid, mem_quality = ( entry.transcript_id,
entry.mQueryCoverage,
entry.mPid,
entry.mQuality )
mem_seq = peptides[mem_id]
mem_extended_seq = extended_peptides[mem_id]
if options.loglevel >= 4:
options.stdlog.write( "# processing: id=%s class=%s\n" % (mem_id, mem_quality))
if mem_id in eliminated_predictions: continue
if mem_extended_seq == rep_extended_seq:
eliminated_predictions[mem_id] = rep_id
eliminated.append( (mem_id, "i") )
elif mem_extended_seq in rep_extended_seq:
eliminated_predictions[mem_id] = rep_id
eliminated.append( (mem_id, "p") )
else:
if mem_quality != this_quality or \
mem_quality in options.quality_exclude_same:
seq1 = alignlib_lite.makeSequence( str(rep_seq) )
seq2 = alignlib_lite.makeSequence( str(mem_seq) )
alignator.align( result, seq1, seq2 )
if options.loglevel >= 5:
options.stdlog.write( "# ali\n%s\n" % alignlib_lite.AlignmentFormatExplicit( result, seq1, seq2 ) )
pidentity = 100 * alignlib_lite.calculatePercentIdentity( result, seq1, seq2 )
num_gaps = result.getNumGaps()
if options.loglevel >= 4:
options.stdlog.write( "# processing: id=%s class=%s pid=%5.2f rep_cov=%i mem_cov=%i\n" %\
( mem_id, mem_quality, pidentity, rep_coverage, mem_coverage ) )
if pidentity >= options.min_identity:
keep = False
if rep_coverage < mem_coverage - options.safety_coverage or \
rep_pid < mem_pid - options.safety_pide:
keep = True
reason = "covpid"
elif num_gaps >= options.max_gaps and \
mem_coverage > rep_coverage - options.safety_coverage:
keep = True
reason = "gaps"
elif mem_coverage >= rep_coverage - options.safety_coverage and \
100 * (result.getColTo() - result.getColFrom()) / len(mem_seq) < options.max_member_coverage:
keep = True
reason = "memcov"
if keep:
options.stdlog.write( "# WARNING: not removing possibly good prediction: %s: rep = %s, mem = %s, rep_cov=%i, rep_pid=%i, mem_cov=%i, mem_pid=%i\n" %\
(reason, rep_id, mem_id, rep_coverage, rep_pid, mem_coverage, mem_pid) )
else:
eliminated_predictions[mem_id] = rep_id
eliminated.append( (mem_id, "h") )
elif pidentity >= options.min_identity_non_genes and \
this_quality in options.quality_genes and \
mem_quality not in options.quality_genes:
if rep_coverage < mem_coverage - options.safety_coverage or \
rep_pid < mem_pid - options.safety_pide:
options.stdlog.write( "# WARNING: not removing possibly good prediction: rep = %s, mem = %s, rep_cov=%i, rep_pid=%i, mem_cov=%i, mem_pid=%i\n" %\
(rep_id, mem_id, rep_coverage, rep_pid, mem_coverage, mem_pid) )
else:
eliminated_predictions[mem_id] = rep_id
eliminated.append( (mem_id, "l") )
return eliminated
