def main(dir, contigs_file1, contigs_file2, unique_contigs_file):
CreateLog(dir)
sys.stdout.level = LogPriority.warning
unique = ContigCollection().loadFromFasta(
open(unique_contigs_file,
"r"), False).filter(lambda contig: len(contig) > 5000)
aligner = Aligner(DirDistributor(os.path.join(dir, "alignments")))
contigs1 = ContigCollection().loadFromFasta(open(contigs_file1, "r"),
False)
cals1 = list(aligner.overlapAlign(unique.unique(), contigs1))
transfers1, term1, all1 = extract_transfers(contigs1, cals1)
contigs2 = ContigCollection().loadFromFasta(open(contigs_file2, "r"),
False)
cals2 = list(aligner.overlapAlign(unique.unique(), contigs2))
transfers2, term2, all2 = extract_transfers(contigs2, cals2)
missing1 = []
missing2 = []
different = dict()
unresolved1 = []
unresolved2 = []
same = []
for ucontig in list(unique) + [contig.rc for contig in unique]:
uid = ucontig.id
in1 = uid in all1
in2 = uid in all2
if not in1 and not in2:
continue
if not in1:
missing1.append(uid)
elif not in2:
missing2.append(uid)
else:
if all1[uid][0] == all2[uid][0]:
same.append(uid)
elif uid in transfers1 and uid in transfers2:
different[uid] = (all1[uid][0], all2[uid][0])
elif uid in transfers1:
unresolved2.append(uid)
elif uid in transfers2:
unresolved1.append(uid)
out = open(os.path.join(dir, "contigs.txt"), "w")
out.write("Different: " + str(different) + "\n")
out.write("Unresolved1: " + str(unresolved1) + "\n")
out.write("Unresolved2: " + str(unresolved2) + "\n")
out.write("Same: " + str(same) + "\n")
out.write("Missing1: " + str(missing1) + "\n")
out.write("Missing2: " + str(missing2) + "\n")
out.write("Contig1 transfers: " + str(transfers1) + "\n")
out.write("Contig1 term: " + str(term1) + "\n")
out.write("Contig2 transfers: " + str(transfers2) + "\n")
out.write("Contig2 term: " + str(term2) + "\n")
out.close()
print contigs_file1, contigs_file2
print len(different), len(unresolved1), len(unresolved2), len(
missing1), len(missing2), len(same)
def main(dir, contigs_files, reference_file, unique_contigs_file):
CreateLog(dir)
sys.stdout.level = LogPriority.warning
ref = ContigCollection().loadFromFasta(open(reference_file, "r"), False)
unique = ContigCollection().loadFromFasta(open(unique_contigs_file, "r"), False).filter(lambda contig: len(contig) > 5000)
aligner = Aligner(DirDistributor(os.path.join(dir, "alignments")))
ref_als= list(aligner.overlapAlign(unique.unique(), ref))
ref_transfers, ref_term, all_ref_als = extract_transfers(ref, ref_als)
for uid in ref_term:
ref_transfers[uid] = ref_term[uid]
print "#", "file", "wrong", "unresolved", "correct", "missing"
for i, contigs_file in enumerate(contigs_files):
contigs = ContigCollection().loadFromFasta(open(contigs_file, "r"), False)
contig_als = list(aligner.overlapAlign(unique.unique(), contigs))
contig_transfers, contig_term, all_contig_als = extract_transfers(contigs, contig_als)
missing = []
wrong = dict()
unresolved = []
correct = []
for uid in ref_transfers:
if uid not in contig_transfers and uid not in contig_term:
# print uid, "missing"
missing.append(uid)
elif uid in contig_transfers:
if ref_transfers[uid][0] == contig_transfers[uid][0]:
# print uid, "correct"
correct.append(uid)
else:
# print uid, "wrong", ref_transfers[uid][0].id, contig_transfers[uid][0].id
wrong[uid] = (ref_transfers[uid][0], contig_transfers[uid][0])
else:
if ref_transfers[uid][0] == contig_term[uid][0]:
# print uid, "correct"
correct.append(uid)
else:
# print uid, "unresolved"
unresolved.append(uid)
out = open(os.path.join(dir, "contigs_" + str(i) +".txt"), "w")
out.write("Wrong: " + str(wrong) + "\n")
out.write("Unresolved: " + str(unresolved) + "\n")
out.write("Correct: " + str(correct) + "\n")
out.write("Missing: " + str(missing) + "\n")
out.write("Contig transfers: " + str(contig_transfers) + "\n")
out.write("Contig term: " + str(contig_term) + "\n")
out.write("Ref transfers: " + str(ref_transfers) + "\n")
out.write("Ref als:\n")
for c in all_ref_als:
out.write(str(c) + "\n")
out.write("Contig als:\n")
for c in all_contig_als:
out.write(str(c) + "\n")
out.close()
print "result", i, contigs_file, len(wrong), len(unresolved), len(correct), len(missing)
def main(ref_file, contig_size, rlen, cov, dir):
basic.ensure_dir_existance(dir)
all_contigs = ContigCollection().loadFromFasta(open(ref_file, "r"), False)
contig_file_name = os.path.join(dir, "contigs.fasta")
contig_file = open(contig_file_name, "w")
reads_file_name = os.path.join(dir, "reads.fasta")
reads_file = open(reads_file_name, "w")
for ref in all_contigs.unique():
if len(ref) < contig_size:
continue
SeqIO.write(ref, contig_file, "fasta")
for i in range(0, len(ref), max(1, rlen / cov)):
read = ref.segment(i, min(i + rlen, len(ref))).asNamedSequence()
SeqIO.write(read, reads_file, "fasta")
reads_file.close()
contig_file.close()
print "Done"
print contig_file_name
print reads_file_name
def CreateContigCollection(graph_file, contigs_file, min_cov, aligner, polisher, reads, force_unique, all_unique):
sys.stdout.info("Creating contig collection")
if force_unique is None and not all_unique:
graph = SimpleGraph().ReadDot(graph_file)
graph.FillSeq(contigs_file)
covs = []
for e in graph.e.values():
covs.append((e.len, e.cov))
tmp_cov = []
total = sum(l for c,l in covs) / 2
for l, c in sorted(covs)[::-1]:
if total < 0:
break
tmp_cov.append((l, c))
total -= l
avg_cov = float(sum([l * c for l, c in tmp_cov])) / sum(l for l, c in tmp_cov)
sys.stdout.info("Average coverage determined:", avg_cov)
nonunique = set()
for edge in graph.e.values():
if edge.unique and edge.len < 20000 and len(graph.v[edge.start].out) > 1:
if edge.cov >= min_cov and (edge.cov < 0.8 * avg_cov or edge.len > 40000):
alter = ContigStorage()
for e in graph.v[edge.start].out:
if e != edge:
alter.add(Contig(e.seq, e.id))
for al in aligner.localAlign([Contig(edge.seq, edge.id)], alter):#type: AlignmentPiece
if al.percentIdentity() > 0.98 and (al.seg_from.left < 100 and al.seg_to.left < 100 and len(al) > min(500, edge.len)):
nonunique.add(edge.id)
nonunique.add(basic.Reverse(edge.id))
contigs = ContigCollection()
for edge in graph.e.values():
if basic.isCanonocal(edge.id):
if edge.unique and (edge.len > params.min_isolated_length or len(graph.v[edge.end].out) > 0 or len(graph.v[edge.start].inc) > 0):
if edge.cov >= min_cov and (edge.cov < 1.5 * avg_cov or edge.len > 40000):
if edge.id in nonunique:
sys.stdout.info("Edge removed based on alignment to alternative:", edge.id, edge.cov, edge.len)
else:
contigs.add(Contig(edge.seq, edge.id))
else:
sys.stdout.info("Edge removed based on coverage:", edge.id, edge.cov, edge.len)
elif (edge.len > 100000 and edge.cov < 1.5 * avg_cov) or (edge.len > 40000 and 1.3 * avg_cov > edge.cov > 0.7 * avg_cov):
contigs.add(Contig(edge.seq, edge.id))
sys.stdout.info("Edge added based on length and coverage:", edge.id, edge.cov, edge.len)
elif force_unique is not None:
sys.stdout.info("Using forced unique edge set")
sys.stdout.trace(force_unique)
contigs = ContigCollection().loadFromFile(contigs_file).filter(lambda contig: contig.id in force_unique)
else:
sys.stdout.info("Considering all contigs unique")
contigs = ContigCollection().loadFromFile(contigs_file)
# contigs.loadFromFasta(open(contigs_file, "r"), num_names=True)
# contigs = contigs.filter(lambda contig: contig.id not in nonunique and len(contig) > params.k + 20)
sys.stdout.info("Created", len(contigs), "initial contigs")
if not all_unique or force_unique is not None:
sys.stdout.info("Polishing contigs")
polished_contigs = polisher.polishMany(reads, list(contigs.unique()))
contigs = ContigCollection().addAll(polished_contigs)
else:
sys.stdout.info("Skipping contig polishing step since manual unique contig initialization was used")
return contigs