def parse_minimap_output(raw_coords_fpath, coords_fpath):
cigar_pattern = re.compile(r'(\d+[M=XIDNSH])')
total_aligned_bases = 0
with open(raw_coords_fpath) as f:
with open(coords_fpath, 'w') as coords_file:
for line in f:
fs = line.split('\t')
if len(fs) < 10:
continue
contig, align_start, align_end, strand, ref_name, ref_start = \
fs[0], fs[2], fs[3], fs[4], fs[5], fs[7]
align_start, align_end, ref_start = map(int, (align_start, align_end, ref_start))
align_start += 1
ref_start += 1
if fs[-1].startswith('cs'):
cs = fs[-1].strip()
cigar = fs[-2]
else:
cs = ''
cigar = fs[-1]
cigar = cigar.split(':')[-1]
strand_direction = 1
if strand == '-':
align_start, align_end = align_end, align_start
strand_direction = -1
align_len = 0
ref_len = 0
matched_bases, bases_in_mapping = map(int, (fs[9], fs[10]))
operations = cigar_pattern.findall(cigar)
for op in operations:
n_bases, operation = int(op[:-1]), op[-1]
if operation == 'S' or operation == 'H':
align_start += n_bases
elif operation == 'M' or operation == '=' or operation == 'X':
align_len += n_bases
ref_len += n_bases
elif operation == 'D':
ref_len += n_bases
elif operation == 'I':
align_len += n_bases
align_end = align_start + (align_len - 1) * strand_direction
ref_end = ref_start + ref_len - 1
total_aligned_bases += align_len
idy = '%.2f' % (matched_bases * 100.0 / bases_in_mapping)
if ref_name != "*":
if float(idy) >= qconfig.min_IDY:
align = Mapping(s1=ref_start, e1=ref_end, s2=align_start, e2=align_end, len1=ref_len,
len2=align_len, idy=idy, ref=ref_name, contig=contig, cigar=cs)
coords_file.write(align.coords_str() + '\n')
else:
split_align(coords_file, align_start, strand_direction, ref_start, ref_name, contig, cs)
def _write_align():
if align_len < qconfig.min_alignment or not ref_len or not align_cs:
return
align_end = align_start + (align_len - 1) * strand_direction
ref_end = ref_start + ref_len - 1
align_idy = '%.2f' % (matched_bases * 100.0 / ref_len)
if float(align_idy) >= qconfig.min_IDY:
align = Mapping(s1=ref_start, e1=ref_end, s2=align_start, e2=align_end, len1=ref_len,
len2=align_len, idy=align_idy, ref=ref_name, contig=contig, cigar=align_cs)
coords_file.write(align.coords_str() + '\n')
def align_and_analyze(is_cyclic,
index,
contigs_fpath,
output_dirpath,
ref_fpath,
reference_chromosomes,
ns_by_chromosomes,
old_contigs_fpath,
bed_fpath,
threads=1):
tmp_output_dirpath = create_minimap_output_dir(output_dirpath)
assembly_label = qutils.label_from_fpath(contigs_fpath)
corr_assembly_label = qutils.label_from_fpath_for_fname(contigs_fpath)
out_basename = join(tmp_output_dirpath, corr_assembly_label)
logger.info(' ' + qutils.index_to_str(index) + assembly_label)
if not qconfig.space_efficient:
log_out_fpath = join(
output_dirpath,
qconfig.contig_report_fname_pattern % corr_assembly_label +
'.stdout')
log_err_fpath = join(
output_dirpath,
qconfig.contig_report_fname_pattern % corr_assembly_label +
'.stderr')
icarus_out_fpath = join(
output_dirpath,
qconfig.icarus_report_fname_pattern % corr_assembly_label)
misassembly_fpath = join(
output_dirpath,
qconfig.contig_report_fname_pattern % corr_assembly_label +
'.mis_contigs.info')
unaligned_info_fpath = join(
output_dirpath,
qconfig.contig_report_fname_pattern % corr_assembly_label +
'.unaligned.info')
else:
log_out_fpath = '/dev/null'
log_err_fpath = '/dev/null'
icarus_out_fpath = '/dev/null'
misassembly_fpath = '/dev/null'
unaligned_info_fpath = '/dev/null'
icarus_out_f = open(icarus_out_fpath, 'w')
icarus_header_cols = [
'S1', 'E1', 'S2', 'E2', 'Reference', 'Contig', 'IDY', 'Ambiguous',
'Best_group'
]
icarus_out_f.write('\t'.join(icarus_header_cols) + '\n')
misassembly_f = open(misassembly_fpath, 'w')
if not qconfig.space_efficient:
logger.info(' ' + qutils.index_to_str(index) + 'Logging to files ' +
log_out_fpath + ' and ' + os.path.basename(log_err_fpath) +
'...')
else:
logger.info(' ' + qutils.index_to_str(index) + 'Logging is disabled.')
coords_fpath, coords_filtered_fpath, unaligned_fpath, used_snps_fpath = get_aux_out_fpaths(
out_basename)
status = align_contigs(coords_fpath, out_basename, ref_fpath,
contigs_fpath, old_contigs_fpath, index, threads,
log_out_fpath, log_err_fpath)
if status != AlignerStatus.OK:
with open(log_err_fpath, 'a') as log_err_f:
if status == AlignerStatus.ERROR:
logger.error(
' ' + qutils.index_to_str(index) +
'Failed aligning contigs ' +
qutils.label_from_fpath(contigs_fpath) +
' to the reference (non-zero exit code). ' +
('Run with the --debug flag to see additional information.'
if not qconfig.debug else ''))
elif status == AlignerStatus.FAILED:
log_err_f.write(
qutils.index_to_str(index) + 'Alignment failed for ' +
contigs_fpath + ':' + coords_fpath + 'doesn\'t exist.\n')
logger.info(' ' + qutils.index_to_str(index) +
'Alignment failed for ' + '\'' + assembly_label +
'\'.')
elif status == AlignerStatus.NOT_ALIGNED:
log_err_f.write(
qutils.index_to_str(index) + 'Nothing aligned for ' +
contigs_fpath + '\n')
logger.info(' ' + qutils.index_to_str(index) +
'Nothing aligned for ' + '\'' + assembly_label +
'\'.')
return status, {}, [], [], []
log_out_f = open(log_out_fpath, 'a')
# Loading the alignment files
log_out_f.write('Parsing coords...\n')
aligns = {}
with open(coords_fpath) as coords_file:
for line in coords_file:
mapping = Mapping.from_line(line)
aligns.setdefault(mapping.contig, []).append(mapping)
# Loading the reference sequences
log_out_f.write('Loading reference...\n') # TODO: move up
ref_features = {}
# Loading the regions (if any)
regions = {}
total_reg_len = 0
total_regions = 0
# # TODO: gff
# log_out_f.write('Loading regions...\n')
# log_out_f.write('\tNo regions given, using whole reference.\n')
for name, seq_len in reference_chromosomes.items():
log_out_f.write('\tLoaded [%s]\n' % name)
regions.setdefault(name, []).append([1, seq_len])
total_regions += 1
total_reg_len += seq_len
log_out_f.write('\tTotal Regions: %d\n' % total_regions)
log_out_f.write('\tTotal Region Length: %d\n' % total_reg_len)
ca_output = CAOutput(stdout_f=log_out_f,
misassembly_f=misassembly_f,
coords_filtered_f=open(coords_filtered_fpath, 'w'),
icarus_out_f=icarus_out_f)
log_out_f.write('Analyzing contigs...\n')
result, ref_aligns, total_indels_info, aligned_lengths, misassembled_contigs, misassemblies_in_contigs, aligned_lengths_by_contigs =\
analyze_contigs(ca_output, contigs_fpath, unaligned_fpath, unaligned_info_fpath, aligns, ref_features, reference_chromosomes, is_cyclic)
log_out_f.write('Analyzing coverage...\n')
if qconfig.show_snps:
log_out_f.write('Writing SNPs into ' + used_snps_fpath + '\n')
total_aligned_bases, indels_info = analyze_coverage(
ref_aligns, reference_chromosomes, ns_by_chromosomes, used_snps_fpath)
total_indels_info += indels_info
cov_stats = {
'SNPs': total_indels_info.mismatches,
'indels_list': total_indels_info.indels_list,
'total_aligned_bases': total_aligned_bases
}
result.update(cov_stats)
result = print_results(contigs_fpath, log_out_f, used_snps_fpath,
total_indels_info, result)
if not qconfig.space_efficient:
## outputting misassembled contigs to separate file
fasta = [(name, seq)
for name, seq in fastaparser.read_fasta(contigs_fpath)
if name in misassembled_contigs.keys()]
fastaparser.write_fasta(
join(output_dirpath,
qutils.name_from_fpath(contigs_fpath) + '.mis_contigs.fa'),
fasta)
if qconfig.is_combined_ref:
alignment_tsv_fpath = join(
output_dirpath, "alignments_" + corr_assembly_label + '.tsv')
unique_contigs_fpath = join(
output_dirpath,
qconfig.unique_contigs_fname_pattern % corr_assembly_label)
logger.debug(' ' + qutils.index_to_str(index) + 'Alignments: ' +
qutils.relpath(alignment_tsv_fpath))
used_contigs = set()
with open(unique_contigs_fpath, 'w') as unique_contigs_f:
with open(alignment_tsv_fpath, 'w') as alignment_tsv_f:
for chr_name, aligns in ref_aligns.items():
alignment_tsv_f.write(chr_name)
contigs = set([align.contig for align in aligns])
for contig in contigs:
alignment_tsv_f.write('\t' + contig)
if qconfig.is_combined_ref:
ref_name = ref_labels_by_chromosomes[chr_name]
align_by_contigs = defaultdict(int)
for align in aligns:
align_by_contigs[align.contig] += align.len2
for contig, aligned_len in align_by_contigs.items():
if contig in used_contigs:
continue
used_contigs.add(contig)
len_cov_pattern = re.compile(
r'_length_([\d\.]+)_cov_([\d\.]+)')
if len_cov_pattern.findall(contig):
contig_len = len_cov_pattern.findall(
contig)[0][0]
contig_cov = len_cov_pattern.findall(
contig)[0][1]
if aligned_len / float(contig_len) > 0.9:
unique_contigs_f.write(ref_name + '\t' +
str(aligned_len) +
'\t' + contig_cov +
'\n')
alignment_tsv_f.write('\n')
close_handlers(ca_output)
logger.info(' ' + qutils.index_to_str(index) + 'Analysis is finished.')
logger.debug('')
if not ref_aligns:
return AlignerStatus.NOT_ALIGNED, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
else:
return AlignerStatus.OK, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
def align_and_analyze(is_cyclic, index, contigs_fpath, output_dirpath, ref_fpath,
old_contigs_fpath, bed_fpath, parallel_by_chr=False, threads=1):
nucmer_output_dirpath = create_nucmer_output_dir(output_dirpath)
assembly_label = qutils.label_from_fpath(contigs_fpath)
corr_assembly_label = qutils.label_from_fpath_for_fname(contigs_fpath)
nucmer_fpath = join(nucmer_output_dirpath, corr_assembly_label)
logger.info(' ' + qutils.index_to_str(index) + assembly_label)
if not qconfig.space_efficient:
log_out_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.stdout')
log_err_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.stderr')
icarus_out_fpath = join(output_dirpath, qconfig.icarus_report_fname_pattern % corr_assembly_label)
misassembly_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.mis_contigs.info')
unaligned_info_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.unaligned.info')
else:
log_out_fpath = '/dev/null'
log_err_fpath = '/dev/null'
icarus_out_fpath = '/dev/null'
misassembly_fpath = '/dev/null'
unaligned_info_fpath = '/dev/null'
icarus_out_f = open(icarus_out_fpath, 'w')
icarus_header_cols = ['S1', 'E1', 'S2', 'E2', 'Reference', 'Contig', 'IDY', 'Ambiguous', 'Best_group']
icarus_out_f.write('\t'.join(icarus_header_cols) + '\n')
misassembly_f = open(misassembly_fpath, 'w')
if not qconfig.space_efficient:
logger.info(' ' + qutils.index_to_str(index) + 'Logging to files ' + log_out_fpath +
' and ' + os.path.basename(log_err_fpath) + '...')
else:
logger.info(' ' + qutils.index_to_str(index) + 'Logging is disabled.')
coords_fpath, coords_filtered_fpath, unaligned_fpath, show_snps_fpath, used_snps_fpath = \
get_nucmer_aux_out_fpaths(nucmer_fpath)
nucmer_status = align_contigs(nucmer_fpath, ref_fpath, contigs_fpath, old_contigs_fpath, index,
parallel_by_chr, threads, log_out_fpath, log_err_fpath)
if nucmer_status != NucmerStatus.OK:
with open(log_err_fpath, 'a') as log_err_f:
if nucmer_status == NucmerStatus.ERROR:
logger.error(' ' + qutils.index_to_str(index) +
'Failed aligning contigs ' + qutils.label_from_fpath(contigs_fpath) +
' to the reference (non-zero exit code). ' +
('Run with the --debug flag to see additional information.' if not qconfig.debug else ''))
elif nucmer_status == NucmerStatus.FAILED:
log_err_f.write(qutils.index_to_str(index) + 'Alignment failed for ' + contigs_fpath + ':' + coords_fpath + 'doesn\'t exist.\n')
logger.info(' ' + qutils.index_to_str(index) + 'Alignment failed for ' + '\'' + assembly_label + '\'.')
elif nucmer_status == NucmerStatus.NOT_ALIGNED:
log_err_f.write(qutils.index_to_str(index) + 'Nothing aligned for ' + contigs_fpath + '\n')
logger.info(' ' + qutils.index_to_str(index) + 'Nothing aligned for ' + '\'' + assembly_label + '\'.')
clean_tmp_files(nucmer_fpath)
return nucmer_status, {}, [], [], []
log_out_f = open(log_out_fpath, 'a')
# Loading the alignment files
log_out_f.write('Parsing coords...\n')
aligns = {}
coords_file = open(coords_fpath)
coords_filtered_file = open(coords_filtered_fpath, 'w')
coords_filtered_file.write(coords_file.readline())
coords_filtered_file.write(coords_file.readline())
for line in coords_file:
if line.strip() == '':
break
assert line[0] != '='
#Clear leading spaces from nucmer output
#Store nucmer lines in an array
mapping = Mapping.from_line(line)
aligns.setdefault(mapping.contig, []).append(mapping)
# Loading the reference sequences
log_out_f.write('Loading reference...\n') # TODO: move up
ref_lens = {}
ref_features = {}
for name, seq in fastaparser.read_fasta(ref_fpath):
name = name.split()[0] # no spaces in reference header
ref_lens[name] = len(seq)
log_out_f.write('\tLoaded [%s]\n' % name)
#Loading the SNP calls
if qconfig.show_snps:
log_out_f.write('Loading SNPs...\n')
used_snps_file = None
snps = {}
if qconfig.show_snps:
prev_line = None
for line in open_gzipsafe(show_snps_fpath):
#print "$line";
line = line.split()
if not line[0].isdigit():
continue
if prev_line and line == prev_line:
continue
ref = line[10]
ctg = line[11]
pos = int(line[0]) # Kolya: python don't convert int<->str types automatically
loc = int(line[3]) # Kolya: same as above
# if (! exists $line[11]) { die "Malformed line in SNP file. Please check that show-snps has completed succesfully.\n$line\n[$line[9]][$line[10]][$line[11]]\n"; }
if pos in snps.setdefault(ref, {}).setdefault(ctg, {}):
snps.setdefault(ref, {}).setdefault(ctg, {})[pos].append(SNP(ref_pos=pos, ctg_pos=loc, ref_nucl=line[1], ctg_nucl=line[2]))
else:
snps.setdefault(ref, {}).setdefault(ctg, {})[pos] = [SNP(ref_pos=pos, ctg_pos=loc, ref_nucl=line[1], ctg_nucl=line[2])]
prev_line = line
used_snps_file = open_gzipsafe(used_snps_fpath, 'w')
# Loading the regions (if any)
regions = {}
total_reg_len = 0
total_regions = 0
# # TODO: gff
# log_out_f.write('Loading regions...\n')
# log_out_f.write('\tNo regions given, using whole reference.\n')
for name, seq_len in ref_lens.items():
regions.setdefault(name, []).append([1, seq_len])
total_regions += 1
total_reg_len += seq_len
log_out_f.write('\tTotal Regions: %d\n' % total_regions)
log_out_f.write('\tTotal Region Length: %d\n' % total_reg_len)
ca_output = CAOutput(stdout_f=log_out_f, misassembly_f=misassembly_f, coords_filtered_f=coords_filtered_file,
used_snps_f=used_snps_file, icarus_out_f=icarus_out_f)
log_out_f.write('Analyzing contigs...\n')
result, ref_aligns, total_indels_info, aligned_lengths, misassembled_contigs, misassemblies_in_contigs, aligned_lengths_by_contigs =\
analyze_contigs(ca_output, contigs_fpath, unaligned_fpath, unaligned_info_fpath, aligns, ref_features, ref_lens, is_cyclic)
# if qconfig.large_genome:
# log_out_f.write('Analyzing large blocks...\n')
# large_misassembly_fpath = add_suffix(misassembly_fpath, 'large_blocks') if not qconfig.space_efficient else '/dev/null'
# ca_large_output = CAOutput(stdout_f=log_out_f, misassembly_f=open(large_misassembly_fpath, 'w'),
# coords_filtered_f=coords_filtered_file, used_snps_f=open('/dev/null', 'w'), icarus_out_f=open('/dev/null', 'w'))
# min_alignment, extensive_mis_threshold = qconfig.min_alignment, qconfig.extensive_misassembly_threshold
# qconfig.min_alignment, qconfig.extensive_misassembly_threshold = qconfig.LARGE_MIN_ALIGNMENT, qconfig.LARGE_EXTENSIVE_MIS_THRESHOLD
# result.update(analyze_contigs(ca_large_output, contigs_fpath, '/dev/null', '/dev/null',
# aligns, ref_features, ref_lens, is_cyclic, large_misassemblies_search=True)[0])
# qconfig.min_alignment, qconfig.extensive_misassembly_threshold = min_alignment, extensive_mis_threshold
log_out_f.write('Analyzing coverage...\n')
if qconfig.show_snps:
log_out_f.write('Writing SNPs into ' + used_snps_fpath + '\n')
result.update(analyze_coverage(ca_output, regions, ref_aligns, ref_features, snps, total_indels_info))
result = print_results(contigs_fpath, log_out_f, used_snps_fpath, total_indels_info, result)
if not qconfig.space_efficient:
## outputting misassembled contigs to separate file
fasta = [(name, seq) for name, seq in fastaparser.read_fasta(contigs_fpath)
if name in misassembled_contigs.keys()]
fastaparser.write_fasta(join(output_dirpath, qutils.name_from_fpath(contigs_fpath) + '.mis_contigs.fa'), fasta)
if qconfig.is_combined_ref:
alignment_tsv_fpath = join(output_dirpath, "alignments_" + corr_assembly_label + '.tsv')
unique_contigs_fpath = join(output_dirpath, qconfig.unique_contigs_fname_pattern % corr_assembly_label)
logger.debug(' ' + qutils.index_to_str(index) + 'Alignments: ' + qutils.relpath(alignment_tsv_fpath))
used_contigs = set()
with open(unique_contigs_fpath, 'w') as unique_contigs_f:
with open(alignment_tsv_fpath, 'w') as alignment_tsv_f:
for chr_name, aligns in ref_aligns.items():
alignment_tsv_f.write(chr_name)
contigs = set([align.contig for align in aligns])
for contig in contigs:
alignment_tsv_f.write('\t' + contig)
if qconfig.is_combined_ref:
ref_name = ref_labels_by_chromosomes[chr_name]
align_by_contigs = defaultdict(int)
for align in aligns:
align_by_contigs[align.contig] += align.len2
for contig, aligned_len in align_by_contigs.items():
if contig in used_contigs:
continue
used_contigs.add(contig)
len_cov_pattern = re.compile(r'_length_([\d\.]+)_cov_([\d\.]+)')
if len_cov_pattern.findall(contig):
contig_len = len_cov_pattern.findall(contig)[0][0]
contig_cov = len_cov_pattern.findall(contig)[0][1]
if aligned_len / float(contig_len) > 0.9:
unique_contigs_f.write(ref_name + '\t' + str(aligned_len) + '\t' + contig_cov + '\n')
alignment_tsv_f.write('\n')
close_handlers(ca_output)
logger.info(' ' + qutils.index_to_str(index) + 'Analysis is finished.')
logger.debug('')
clean_tmp_files(nucmer_fpath)
if not qconfig.no_gzip:
compress_nucmer_output(logger, nucmer_fpath)
if not ref_aligns:
return NucmerStatus.NOT_ALIGNED, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
else:
return NucmerStatus.OK, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
def align_and_analyze(is_cyclic, index, contigs_fpath, output_dirpath, ref_fpath,
old_contigs_fpath, bed_fpath, parallel_by_chr=False, threads=1):
nucmer_output_dirpath = create_nucmer_output_dir(output_dirpath)
assembly_label = qutils.label_from_fpath(contigs_fpath)
corr_assembly_label = qutils.label_from_fpath_for_fname(contigs_fpath)
nucmer_fpath = join(nucmer_output_dirpath, corr_assembly_label)
logger.info(' ' + qutils.index_to_str(index) + assembly_label)
if not qconfig.space_efficient:
log_out_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.stdout')
log_err_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.stderr')
icarus_out_fpath = join(output_dirpath, qconfig.icarus_report_fname_pattern % corr_assembly_label)
misassembly_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.mis_contigs.info')
unaligned_info_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.unaligned.info')
else:
log_out_fpath = '/dev/null'
log_err_fpath = '/dev/null'
icarus_out_fpath = '/dev/null'
misassembly_fpath = '/dev/null'
unaligned_info_fpath = '/dev/null'
icarus_out_f = open(icarus_out_fpath, 'w')
icarus_header_cols = ['S1', 'E1', 'S2', 'E2', 'Reference', 'Contig', 'IDY', 'Ambiguous', 'Best_group']
icarus_out_f.write('\t'.join(icarus_header_cols) + '\n')
misassembly_f = open(misassembly_fpath, 'w')
if not qconfig.space_efficient:
logger.info(' ' + qutils.index_to_str(index) + 'Logging to files ' + log_out_fpath +
' and ' + os.path.basename(log_err_fpath) + '...')
else:
logger.info(' ' + qutils.index_to_str(index) + 'Logging is disabled.')
coords_fpath, coords_filtered_fpath, unaligned_fpath, show_snps_fpath, used_snps_fpath = \
get_nucmer_aux_out_fpaths(nucmer_fpath)
nucmer_status = align_contigs(nucmer_fpath, ref_fpath, contigs_fpath, old_contigs_fpath, index,
parallel_by_chr, threads, log_out_fpath, log_err_fpath)
if nucmer_status != NucmerStatus.OK:
with open(log_err_fpath, 'a') as log_err_f:
if nucmer_status == NucmerStatus.ERROR:
logger.error(' ' + qutils.index_to_str(index) +
'Failed aligning contigs ' + qutils.label_from_fpath(contigs_fpath) +
' to the reference (non-zero exit code). ' +
('Run with the --debug flag to see additional information.' if not qconfig.debug else ''))
elif nucmer_status == NucmerStatus.FAILED:
log_err_f.write(qutils.index_to_str(index) + 'Alignment failed for ' + contigs_fpath + ':' + coords_fpath + 'doesn\'t exist.\n')
logger.info(' ' + qutils.index_to_str(index) + 'Alignment failed for ' + '\'' + assembly_label + '\'.')
elif nucmer_status == NucmerStatus.NOT_ALIGNED:
log_err_f.write(qutils.index_to_str(index) + 'Nothing aligned for ' + contigs_fpath + '\n')
logger.info(' ' + qutils.index_to_str(index) + 'Nothing aligned for ' + '\'' + assembly_label + '\'.')
clean_tmp_files(nucmer_fpath)
return nucmer_status, {}, [], [], []
log_out_f = open(log_out_fpath, 'a')
# Loading the alignment files
log_out_f.write('Parsing coords...\n')
aligns = {}
coords_file = open(coords_fpath)
coords_filtered_file = open(coords_filtered_fpath, 'w')
coords_filtered_file.write(coords_file.readline())
coords_filtered_file.write(coords_file.readline())
for line in coords_file:
if line.strip() == '':
break
assert line[0] != '='
#Clear leading spaces from nucmer output
#Store nucmer lines in an array
mapping = Mapping.from_line(line)
aligns.setdefault(mapping.contig, []).append(mapping)
# Loading the reference sequences
log_out_f.write('Loading reference...\n') # TODO: move up
references = {}
ref_features = {}
for name, seq in fastaparser.read_fasta(ref_fpath):
name = name.split()[0] # no spaces in reference header
references[name] = seq
log_out_f.write('\tLoaded [%s]\n' % name)
#Loading the SNP calls
if qconfig.show_snps:
log_out_f.write('Loading SNPs...\n')
used_snps_file = None
snps = {}
if qconfig.show_snps:
prev_line = None
for line in open_gzipsafe(show_snps_fpath):
#print "$line";
line = line.split()
if not line[0].isdigit():
continue
if prev_line and line == prev_line:
continue
ref = line[10]
ctg = line[11]
pos = int(line[0]) # Kolya: python don't convert int<->str types automatically
loc = int(line[3]) # Kolya: same as above
# if (! exists $line[11]) { die "Malformed line in SNP file. Please check that show-snps has completed succesfully.\n$line\n[$line[9]][$line[10]][$line[11]]\n"; }
if pos in snps.setdefault(ref, {}).setdefault(ctg, {}):
snps.setdefault(ref, {}).setdefault(ctg, {})[pos].append(SNP(ref_pos=pos, ctg_pos=loc, ref_nucl=line[1], ctg_nucl=line[2]))
else:
snps.setdefault(ref, {}).setdefault(ctg, {})[pos] = [SNP(ref_pos=pos, ctg_pos=loc, ref_nucl=line[1], ctg_nucl=line[2])]
prev_line = line
used_snps_file = open_gzipsafe(used_snps_fpath, 'w')
# Loading the regions (if any)
regions = {}
ref_lens = {}
total_reg_len = 0
total_regions = 0
# # TODO: gff
# log_out_f.write('Loading regions...\n')
# log_out_f.write('\tNo regions given, using whole reference.\n')
for name, seq in references.items():
regions.setdefault(name, []).append([1, len(seq)])
ref_lens[name] = len(seq)
total_regions += 1
total_reg_len += ref_lens[name]
log_out_f.write('\tTotal Regions: %d\n' % total_regions)
log_out_f.write('\tTotal Region Length: %d\n' % total_reg_len)
ca_output = CAOutput(stdout_f=log_out_f, misassembly_f=misassembly_f, coords_filtered_f=coords_filtered_file,
used_snps_f=used_snps_file, icarus_out_f=icarus_out_f)
log_out_f.write('Analyzing contigs...\n')
result, ref_aligns, total_indels_info, aligned_lengths, misassembled_contigs, misassemblies_in_contigs, aligned_lengths_by_contigs =\
analyze_contigs(ca_output, contigs_fpath, unaligned_fpath, unaligned_info_fpath, aligns, ref_features, ref_lens, is_cyclic)
log_out_f.write('Analyzing coverage...\n')
if qconfig.show_snps:
log_out_f.write('Writing SNPs into ' + used_snps_fpath + '\n')
result.update(analyze_coverage(ca_output, regions, ref_aligns, ref_features, snps, total_indels_info))
result = print_results(contigs_fpath, log_out_f, used_snps_fpath, total_indels_info, result)
if not qconfig.space_efficient:
## outputting misassembled contigs to separate file
fasta = [(name, seq) for name, seq in fastaparser.read_fasta(contigs_fpath)
if name in misassembled_contigs.keys()]
fastaparser.write_fasta(join(output_dirpath, qutils.name_from_fpath(contigs_fpath) + '.mis_contigs.fa'), fasta)
if qconfig.is_combined_ref:
alignment_tsv_fpath = join(output_dirpath, "alignments_" + corr_assembly_label + '.tsv')
unique_contigs_fpath = join(output_dirpath, qconfig.unique_contigs_fname_pattern % corr_assembly_label)
logger.debug(' ' + qutils.index_to_str(index) + 'Alignments: ' + qutils.relpath(alignment_tsv_fpath))
used_contigs = set()
with open(unique_contigs_fpath, 'w') as unique_contigs_f:
with open(alignment_tsv_fpath, 'w') as alignment_tsv_f:
for chr_name, aligns in ref_aligns.items():
alignment_tsv_f.write(chr_name)
contigs = set([align.contig for align in aligns])
for contig in contigs:
alignment_tsv_f.write('\t' + contig)
if qconfig.is_combined_ref:
ref_name = ref_labels_by_chromosomes[chr_name]
align_by_contigs = defaultdict(int)
for align in aligns:
align_by_contigs[align.contig] += align.len2
for contig, aligned_len in align_by_contigs.items():
if contig in used_contigs:
continue
used_contigs.add(contig)
len_cov_pattern = re.compile(r'_length_([\d\.]+)_cov_([\d\.]+)')
if len_cov_pattern.findall(contig):
contig_len = len_cov_pattern.findall(contig)[0][0]
contig_cov = len_cov_pattern.findall(contig)[0][1]
if aligned_len / float(contig_len) > 0.9:
unique_contigs_f.write(ref_name + '\t' + str(aligned_len) + '\t' + contig_cov + '\n')
alignment_tsv_f.write('\n')
close_handlers(ca_output)
logger.info(' ' + qutils.index_to_str(index) + 'Analysis is finished.')
logger.debug('')
clean_tmp_files(nucmer_fpath)
if not qconfig.no_gzip:
compress_nucmer_output(logger, nucmer_fpath)
if not ref_aligns:
return NucmerStatus.NOT_ALIGNED, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
else:
return NucmerStatus.OK, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
def split_align(coords_file, align_start, strand_direction, ref_start,
ref_name, contig, cs):
def _write_align():
if align.len2 < qconfig.min_alignment or not align.len1 or not align.cigar:
return
align.e1 = align.s1 + align.len1 - 1
align.e2 = align.s2 + (align.len2 - 1) * strand_direction
align.idy = '%.2f' % (matched_bases * 100.0 /
max(align.len1, align.len2))
if float(align.idy) >= qconfig.min_IDY:
coords_file.write(align.coords_str() + '\n')
def _try_split(matched_bases, prev_op, n_refbases=0, n_alignbases=0):
## split alignment in positions of indels or stretch of mismatches to get smaller alignments with higher identity
if n_alignbases > SPLIT_ALIGN_THRESHOLD or n_refbases > SPLIT_ALIGN_THRESHOLD:
_write_align()
align.s1 += align.len1 + n_refbases
align.s2 += (align.len2 + n_alignbases) * strand_direction
align.len1, align.len2 = 0, 0
align.cigar = ''
matched_bases = 0
else:
align.len1 += n_refbases
align.len2 += n_alignbases
align.cigar += prev_op
return matched_bases
matched_bases = 0
align = Mapping(s1=ref_start,
e1=ref_start,
s2=align_start,
e2=align_start,
len1=0,
len2=0,
ref=ref_name,
contig=contig,
cigar='')
cur_mismatch_stretch = ''
for op in parse_cs_tag(cs):
if op.startswith('*'):
cur_mismatch_stretch += op
continue
if cur_mismatch_stretch:
n_bases = cur_mismatch_stretch.count('*')
matched_bases = _try_split(matched_bases, cur_mismatch_stretch,
n_bases, n_bases)
cur_mismatch_stretch = ''
if op.startswith(':'):
n_bases = int(op[1:])
align.cigar += op
align.len1 += n_bases
align.len2 += n_bases
matched_bases += n_bases
else:
n_bases = len(op) - 1
if op.startswith('+'):
matched_bases = _try_split(matched_bases,
op,
n_alignbases=n_bases)
elif op.startswith('-'):
matched_bases = _try_split(matched_bases,
op,
n_refbases=n_bases)
_write_align()
def align_and_analyze(is_cyclic, index, contigs_fpath, output_dirpath, ref_fpath,
reference_chromosomes, ns_by_chromosomes, old_contigs_fpath, bed_fpath, threads=1):
tmp_output_dirpath = create_minimap_output_dir(output_dirpath)
assembly_label = qutils.label_from_fpath(contigs_fpath)
corr_assembly_label = qutils.label_from_fpath_for_fname(contigs_fpath)
out_basename = join(tmp_output_dirpath, corr_assembly_label)
logger.info(' ' + qutils.index_to_str(index) + assembly_label)
if not qconfig.space_efficient:
log_out_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.stdout')
log_err_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.stderr')
icarus_out_fpath = join(output_dirpath, qconfig.icarus_report_fname_pattern % corr_assembly_label)
misassembly_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.mis_contigs.info')
unaligned_info_fpath = join(output_dirpath, qconfig.contig_report_fname_pattern % corr_assembly_label + '.unaligned.info')
else:
log_out_fpath = '/dev/null'
log_err_fpath = '/dev/null'
icarus_out_fpath = '/dev/null'
misassembly_fpath = '/dev/null'
unaligned_info_fpath = '/dev/null'
icarus_out_f = open(icarus_out_fpath, 'w')
icarus_header_cols = ['S1', 'E1', 'S2', 'E2', 'Reference', 'Contig', 'IDY', 'Ambiguous', 'Best_group']
icarus_out_f.write('\t'.join(icarus_header_cols) + '\n')
misassembly_f = open(misassembly_fpath, 'w')
if not qconfig.space_efficient:
logger.info(' ' + qutils.index_to_str(index) + 'Logging to files ' + log_out_fpath +
' and ' + os.path.basename(log_err_fpath) + '...')
else:
logger.info(' ' + qutils.index_to_str(index) + 'Logging is disabled.')
coords_fpath, coords_filtered_fpath, unaligned_fpath, used_snps_fpath = get_aux_out_fpaths(out_basename)
status = align_contigs(coords_fpath, out_basename, ref_fpath, contigs_fpath, old_contigs_fpath, index, threads,
log_out_fpath, log_err_fpath)
if status != AlignerStatus.OK:
with open(log_err_fpath, 'a') as log_err_f:
if status == AlignerStatus.ERROR:
logger.error(' ' + qutils.index_to_str(index) +
'Failed aligning contigs ' + qutils.label_from_fpath(contigs_fpath) +
' to the reference (non-zero exit code). ' +
('Run with the --debug flag to see additional information.' if not qconfig.debug else ''))
elif status == AlignerStatus.FAILED:
log_err_f.write(qutils.index_to_str(index) + 'Alignment failed for ' + contigs_fpath + ':' + coords_fpath + 'doesn\'t exist.\n')
logger.info(' ' + qutils.index_to_str(index) + 'Alignment failed for ' + '\'' + assembly_label + '\'.')
elif status == AlignerStatus.NOT_ALIGNED:
log_err_f.write(qutils.index_to_str(index) + 'Nothing aligned for ' + contigs_fpath + '\n')
logger.info(' ' + qutils.index_to_str(index) + 'Nothing aligned for ' + '\'' + assembly_label + '\'.')
return status, {}, [], [], []
log_out_f = open(log_out_fpath, 'a')
# Loading the alignment files
log_out_f.write('Parsing coords...\n')
aligns = {}
with open(coords_fpath) as coords_file:
for line in coords_file:
mapping = Mapping.from_line(line)
aligns.setdefault(mapping.contig, []).append(mapping)
# Loading the reference sequences
log_out_f.write('Loading reference...\n') # TODO: move up
ref_features = {}
# Loading the regions (if any)
regions = {}
total_reg_len = 0
total_regions = 0
# # TODO: gff
# log_out_f.write('Loading regions...\n')
# log_out_f.write('\tNo regions given, using whole reference.\n')
for name, seq_len in reference_chromosomes.items():
log_out_f.write('\tLoaded [%s]\n' % name)
regions.setdefault(name, []).append([1, seq_len])
total_regions += 1
total_reg_len += seq_len
log_out_f.write('\tTotal Regions: %d\n' % total_regions)
log_out_f.write('\tTotal Region Length: %d\n' % total_reg_len)
ca_output = CAOutput(stdout_f=log_out_f, misassembly_f=misassembly_f, coords_filtered_f=open(coords_filtered_fpath, 'w'),
icarus_out_f=icarus_out_f)
log_out_f.write('Analyzing contigs...\n')
result, ref_aligns, total_indels_info, aligned_lengths, misassembled_contigs, misassemblies_in_contigs, aligned_lengths_by_contigs =\
analyze_contigs(ca_output, contigs_fpath, unaligned_fpath, unaligned_info_fpath, aligns, ref_features, reference_chromosomes, is_cyclic)
log_out_f.write('Analyzing coverage...\n')
if qconfig.show_snps:
log_out_f.write('Writing SNPs into ' + used_snps_fpath + '\n')
total_aligned_bases, indels_info = analyze_coverage(ref_aligns, reference_chromosomes, ns_by_chromosomes, used_snps_fpath)
total_indels_info += indels_info
cov_stats = {'SNPs': total_indels_info.mismatches, 'indels_list': total_indels_info.indels_list, 'total_aligned_bases': total_aligned_bases}
result.update(cov_stats)
result = print_results(contigs_fpath, log_out_f, used_snps_fpath, total_indels_info, result)
if not qconfig.space_efficient:
## outputting misassembled contigs to separate file
fasta = [(name, seq) for name, seq in fastaparser.read_fasta(contigs_fpath)
if name in misassembled_contigs.keys()]
fastaparser.write_fasta(join(output_dirpath, qutils.name_from_fpath(contigs_fpath) + '.mis_contigs.fa'), fasta)
if qconfig.is_combined_ref:
alignment_tsv_fpath = join(output_dirpath, "alignments_" + corr_assembly_label + '.tsv')
unique_contigs_fpath = join(output_dirpath, qconfig.unique_contigs_fname_pattern % corr_assembly_label)
logger.debug(' ' + qutils.index_to_str(index) + 'Alignments: ' + qutils.relpath(alignment_tsv_fpath))
used_contigs = set()
with open(unique_contigs_fpath, 'w') as unique_contigs_f:
with open(alignment_tsv_fpath, 'w') as alignment_tsv_f:
for chr_name, aligns in ref_aligns.items():
alignment_tsv_f.write(chr_name)
contigs = set([align.contig for align in aligns])
for contig in contigs:
alignment_tsv_f.write('\t' + contig)
if qconfig.is_combined_ref:
ref_name = ref_labels_by_chromosomes[chr_name]
align_by_contigs = defaultdict(int)
for align in aligns:
align_by_contigs[align.contig] += align.len2
for contig, aligned_len in align_by_contigs.items():
if contig in used_contigs:
continue
used_contigs.add(contig)
len_cov_pattern = re.compile(r'_length_([\d\.]+)_cov_([\d\.]+)')
if len_cov_pattern.findall(contig):
contig_len = len_cov_pattern.findall(contig)[0][0]
contig_cov = len_cov_pattern.findall(contig)[0][1]
if aligned_len / float(contig_len) > 0.9:
unique_contigs_f.write(ref_name + '\t' + str(aligned_len) + '\t' + contig_cov + '\n')
alignment_tsv_f.write('\n')
close_handlers(ca_output)
logger.info(' ' + qutils.index_to_str(index) + 'Analysis is finished.')
logger.debug('')
if not ref_aligns:
return AlignerStatus.NOT_ALIGNED, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs
else:
return AlignerStatus.OK, result, aligned_lengths, misassemblies_in_contigs, aligned_lengths_by_contigs