def get_transcript_from_gene_pred(in_gene_pred, bam_fn=""):
trans = dd(lambda: [])
ut.err_format_time('get_transcript_from_bed12',
'Loading transcript from {} ... '.format(in_gene_pred))
header_ele = [
'transID', 'chrom', 'strand', 'transStart', 'transEnd', 'cdsStart',
'cdsEnd', 'blockCount', 'exonStarts', 'exonEnds', 'score', 'geneID',
'cdsStartStatus', 'cdsEndStatus', 'exonFrame'
]
pred_header = {header_ele[i]: i for i in range(len(header_ele))}
# bam = ps.AlignmentFile(bam_fn) if bam_fn else None
with open(in_gene_pred, 'r') as pred:
for line in pred:
line = line.rstrip()
if len(line) < 1: continue
if line.startswith('#'): continue
ele = line.rsplit('\t')
# chrom = ele[bed_header['chrom']]
# strand = ele[bed_header['strand']]
start_array, end_array = ele[pred_header['exonStarts']].split(
','), ele[pred_header['exonEnds']].split(',')
gene_id = ele[pred_header['geneID']]
if '' in start_array: start_array.remove('')
if '' in end_array: end_array.remove('')
exon_start = [int(i) + 1 for i in start_array]
exon_end = [int(i) for i in end_array]
# tid = bam.get_tid(chrom) if bam else chrom
# is_rev = strand == '-'
coor = []
for s, e in zip(exon_start, exon_end):
coor.extend([s, e])
trans[gene_id].append('_'.join(map(str, coor)))
ut.err_format_time('get_transcript_from_gene_pred',
'Loading transcript from {} done!'.format(in_gene_pred))
# if bam: bam.close()
return trans
def bam_classify(in_bam_fn,
high_bam_fn,
low_bam_fn,
high_max_ratio=high_max_ratio,
high_min_ratio=high_min_ratio,
high_iden_ratio=high_iden_ratio,
high_repeat_ratio=high_repeat_ratio,
low_repeat_ratio=low_repeat_ratio):
ut.err_format_time('classify_bam_core',
'Processing {} ... '.format(in_bam_fn))
with ps.AlignmentFile(in_bam_fn) as in_bam, ps.AlignmentFile(high_bam_fn, 'wb', template=in_bam) as high_bam, \
ps.AlignmentFile(low_bam_fn, 'wb', template=in_bam) as low_bam:
cnt = 0
r_array = []
r_name = ''
for r in in_bam:
if pb.is_unmapped(r): continue
if r.query_name != r_name:
cla_record(r_array, high_bam, low_bam, high_max_ratio,
high_min_ratio, high_iden_ratio, high_repeat_ratio,
low_repeat_ratio)
r_name = r.query_name
r_array = [r]
else:
r_array.append(r)
cnt += 1
if cnt % 100000 == 0:
ut.err_format_time('classify_bam_core',
'{} BAM records done ... '.format(cnt))
if r_array:
cla_record(r_array, high_bam, low_bam, high_max_ratio,
high_min_ratio, high_iden_ratio, high_repeat_ratio,
low_repeat_ratio)
ut.err_format_time('classify_bam_core',
'Processing {} done.'.format(in_bam_fn))
return high_bam_fn, low_bam_fn
def isocirc_comp_core(args):
a_fn = args.a_input
a_type = args.a_type
a_dict = get_input(a_fn, a_type, args.detailed)
b_fn = args.b_input
b_type = args.b_type
b_dict = get_input(b_fn, b_type, args.detailed)
if args.overlap:
ut.err_format_time('get_overlap', '{} {} ... '.format(a_fn, b_fn))
if args.detailed:
get_detailed_overlap(a_dict, b_dict, args.back_dis, args.inter_dis,
args.overlap_entry, args.overlap)
else:
get_overlap(a_dict, b_dict, args.back_dis, args.overlap_entry,
args.overlap)
ut.err_format_time('get_overlap', '{} {} done!'.format(a_fn, b_fn))
if args.a_only:
ut.err_format_time('get_A_only', '{} {} ... '.format(a_fn, b_fn))
get_only(a_dict, b_dict, args.back_dis, args.a_only)
ut.err_format_time('get_A_only', '{} {} done!'.format(a_fn, b_fn))
if args.b_only:
ut.err_format_time('get_B_only', '{} {} ... '.format(a_fn, b_fn))
get_only(b_dict, a_dict, args.back_dis, args.b_only)
ut.err_format_time('get_B_only', '{} {} done!'.format(a_fn, b_fn))
return
def stats_core(long_read_len, cons_info, cons_bam, isoform_out,
all_bsj_stats_dict, stats_out):
# basic stats of read/cons
tot_read_n, tot_base, tot_read_cons_n, tot_cons_n, tot_cons_base, tot_map_cons_n, tot_map_cons_base = 0, 0, 0, 0, 0, 0, 0
with open(long_read_len, 'r') as in_fp:
for line in in_fp:
[name, read_len] = line.rsplit()
read_len = int(read_len)
tot_read_n += 1
tot_base += read_len
tot_map_read_n, tot_map_cons_n, tot_map_cons_base = get_map_stats(cons_bam)
with open(cons_info) as in_fp:
cons_names = dict()
for line in in_fp:
ele = line.rsplit()
cons_name = ele[0].rsplit('_cons')[0]
cons_names[cons_name] = 1
tot_cons_n += 1
tot_cons_base += int(ele[2])
tot_read_cons_n = len(cons_names)
# detailed stats of circRNA
# tot_known_circRNA
tot_isoform, tot_bsj, tot_known_bsj, tot_circRNA_read_n, tot_iso_with_known_bsj, tot_known_bsj_read_n, tot_iso_with_cano_bsj, tot_cano_bsj_read_n = 0, 0, dd(
lambda: 0), 0, 0, 0, 0, 0
tot_iso_with_cano_sj, tot_read_with_cano_sj = 0, 0
tot_iso_with_high_sj, tot_iso_with_known_ss, tot_iso_with_high_sj_known_ss = 0, 0, 0
tot_read_with_high_sj, tot_read_with_known_ss, tot_read_with_high_sj_known_ss = 0, 0, 0
tot_full_iso, tot_full_read = 0, 0
tot_full_iso_bsj_fsm_iso, tot_full_iso_bsj_fsm_read, tot_full_iso_bsj_nic_iso, tot_full_iso_bsj_nic_read, tot_full_iso_bsj_nnc_iso, tot_full_iso_bsj_nnc_read = 0, 0, 0, 0, 0, 0
tot_full_iso_int_fsm_iso, tot_full_iso_int_fsm_read, tot_full_iso_int_nic_iso, tot_full_iso_int_nic_read, tot_full_iso_int_nnc_iso, tot_full_iso_int_nnc_read = 0, 0, 0, 0, 0, 0
bsj_dict = dict()
iso_dict = dict()
full_iso = dict()
# non_full_iso = dict()
with open(isoform_out) as in_fp:
for line in in_fp:
if line.startswith('#'): continue
ele = line.rsplit()
read_cnt = int(ele[idx['readCount']])
bsj = (ele[idx['chrom']], ele[idx['startCoor0based']],
ele[idx['endCoor']])
if bsj not in bsj_dict:
bsj_dict[bsj] = 1
tot_bsj += 1
for i, known_bsj in enumerate(
ele[idx['isKnownBSJ']].rsplit(',')):
tot_known_bsj[i] += (known_bsj == 'True')
tot_known_bsj[i + 1] += ('False' not in ele[idx['isKnownBSJ']])
tot_circRNA_read_n += read_cnt
iso = (ele[idx['chrom']], ele[idx['startCoor0based']],
ele[idx['endCoor']], ele[idx['blockCount']],
ele[idx['blockSize']], ele[idx['blockStarts']])
# is_full = ele[idx['isFullLength']] == 'True'
# if is_full and iso in non_full_iso:
# print('Full\t{}'.format(ele[0]))
# if not is_full and iso in full_iso:
# print('Non-full\t{}'.format(ele[0]))
# if is_full: full_iso[iso] = 1
# else: non_full_iso[iso] = 1
if iso not in iso_dict:
isoform_inc_cnt = 1
iso_dict[iso] = 1
else:
isoform_inc_cnt = 0
tot_isoform += isoform_inc_cnt
if 'False' not in ele[idx['isKnownSS']]:
tot_read_with_known_ss += read_cnt
tot_iso_with_known_ss += isoform_inc_cnt
# if 'False' not in ele[idx['isCanoSJ']]:
# tot_iso_with_cano_sj += isoform_inc_cnt
# tot_read_with_cano_sj += read_cnt
if 'False' not in ele[idx['isHighFSJ']]:
tot_iso_with_high_sj += isoform_inc_cnt
tot_read_with_high_sj += read_cnt
if 'False' not in ele[idx['isKnownSS']]:
tot_iso_with_high_sj_known_ss += isoform_inc_cnt
tot_read_with_high_sj_known_ss += read_cnt
if ele[idx['isFullLength']] == 'True':
full_iso[iso] = 1
tot_full_iso += isoform_inc_cnt
tot_full_read += read_cnt
if ele[idx['BSJCate']] == 'FSM':
tot_full_iso_bsj_fsm_iso += isoform_inc_cnt
tot_full_iso_bsj_fsm_read += read_cnt
elif ele[idx['BSJCate']] == 'NIC':
tot_full_iso_bsj_nic_iso += isoform_inc_cnt
tot_full_iso_bsj_nic_read += read_cnt
else:
tot_full_iso_bsj_nnc_iso += isoform_inc_cnt
tot_full_iso_bsj_nnc_read += read_cnt
if ele[idx['FSJCate']] == 'FSM':
tot_full_iso_int_fsm_iso += isoform_inc_cnt
tot_full_iso_int_fsm_read += read_cnt
elif ele[idx['FSJCate']] == 'NIC':
tot_full_iso_int_nic_iso += isoform_inc_cnt
tot_full_iso_int_nic_read += read_cnt
else:
tot_full_iso_int_nnc_iso += isoform_inc_cnt
tot_full_iso_int_nnc_read += read_cnt
ut.err_format_time('basic_stats_core', 'Writing basic stats to file ... ')
with open(stats_out, 'w') as out:
out.write('#' + __program__ + '\t' + __version__ + '\n')
out.write(
'# cons=consensus sequence, high=high-confidence, cano=canonical, BSJ=back-splice junction, FSJ=forward-splice junction, SS=splice site\n'
)
out.write('1_Total_reads\t{:,}\n'.format(tot_read_n))
# out.write('Total_base\t{:,}\n'.format(tot_base))
out.write('2_Total_reads_with_cons\t{:,}\n'.format(tot_read_cons_n))
out.write(
'3_Total_mappable_reads_with_cons\t{:,}\n'.format(tot_map_read_n))
out.write('4_Total_reads_with_candidate_BSJs\t{:,}\n'.format(
all_bsj_stats_dict['read_with_bsj_n']))
out.write('5_Total_candidate_BSJs\t{:,}\n'.format(
all_bsj_stats_dict['bsj_n']))
if len(all_bsj_stats_dict['known_bsj_n']) > 2:
for i in all_bsj_stats_dict['known_bsj_n']:
bsj_idx = 'all' if i == len(
all_bsj_stats_dict['known_bsj_n']) - 1 else i
out.write('6_Total_candidate_BSJs_known_in_{}\t{:,}\n'.format(
bsj_idx, all_bsj_stats_dict['known_bsj_n'][i]))
else:
out.write('6_Total_known_candidate_BSJs\t{:,}\n'.format(
all_bsj_stats_dict['known_bsj_n'][0]))
out.write('7_Total_reads_with_high_confidence_BSJs\t{:,}\n'.format(
tot_circRNA_read_n))
out.write('8_Total_high_confidence_BSJs\t{:,}\n'.format(tot_bsj))
if len(tot_known_bsj) > 2:
for i in tot_known_bsj:
bsj_idx = 'all' if i == len(tot_known_bsj) - 1 else i
out.write(
'9_Total_high_confidence_BSJs_known_in_{}\t{:,}\n'.format(
bsj_idx, tot_known_bsj[i]))
else:
out.write('9_Total_known_high_confidence_BSJs\t{:,}\n'.format(
tot_known_bsj[0]))
out.write(
'10_Total_isoforms_with_high_BSJs\t{:,}\n'.format(tot_isoform))
# out.write('11_Total_isoforms_with_high_BSJs_cano_SJs\t{:,}\n'.format(tot_iso_with_cano_sj))
out.write('11_Total_isoforms_with_high_BSJs_high_FSJs\t{:,}\n'.format(
tot_iso_with_high_sj))
out.write('12_Total_isoforms_with_high_BSJ_known_SSs\t{:,}\n'.format(
tot_iso_with_known_ss))
out.write(
'13_Total_isoforms_with_high_BSJs_high_FSJs_known_SSs\t{:,}\n'.
format(tot_iso_with_high_sj_known_ss))
out.write('14_Total_full_length_isoforms\t{:,}\n'.format(
len(full_iso))) #tot_full_iso))
out.write('15_Total_reads_for_full_length_isoforms\t{:,}\n'.format(
tot_full_read))
# FSM/NIC/NNC
out.write('16_Total_full_length_isoforms_with_FSM_BSJ\t{:,}\n'.format(
tot_full_iso_bsj_fsm_iso))
out.write(
'17_Total_reads_for_full_length_isoforms_with_FSM_BSJ\t{:,}\n'.
format(tot_full_iso_bsj_fsm_read))
out.write('18_Total_full_length_isoforms_with_NIC_BSJ\t{:,}\n'.format(
tot_full_iso_bsj_nic_iso))
out.write(
'19_Total_reads_for_full_length_isoforms_with_NIC_BSJ\t{:,}\n'.
format(tot_full_iso_bsj_nic_read))
out.write('20_Total_full_length_isoforms_with_NNC_BSJ\t{:,}\n'.format(
tot_full_iso_bsj_nnc_iso))
out.write(
'21_Total_reads_for_full_length_isoforms_with_NNC_BSJ\t{:,}\n'.
format(tot_full_iso_bsj_nnc_read))
out.write('22_Total_full_length_isoform_with_FSM_FSJ\t{:,}\n'.format(
tot_full_iso_int_fsm_iso))
out.write(
'23_Total_reads_full_length_isoforms_with_FSM_FSJ\t{:,}\n'.format(
tot_full_iso_int_fsm_read))
out.write('24_Total_full_length_isoforms_with_NIC_FSJ\t{:,}\n'.format(
tot_full_iso_int_nic_iso))
out.write(
'25_Total_reads_for_full_length_isoforms_with_NIC_FSJ\t{:,}\n'.
format(tot_full_iso_int_nic_read))
out.write('26_Total_full_length_isoforms_with_NNC_FSJ\t{:,}\n'.format(
tot_full_iso_int_nnc_iso))
out.write(
'27_Total_reads_for_full_length_isoforms_with_NNC_FSJ\t{:,}\n'.
format(tot_full_iso_int_nnc_read))
ut.err_format_time('basic_stats_core', 'Writing basic stats to file done!')
def restore_gff_db(gtf_fn):
gtf_db = None
if gtf_fn is not None:
gtf_db_fn = gtf_fn + '.gffdb'
if not os.path.isfile(gtf_db_fn):
try:
# check if 'gene' or 'transcript' is in GTF
disable_gene, disable_trans = False, False
with open(gtf_fn) as fp:
l = 0
for line in fp:
line = line.rstrip()
if len(line) < 1: continue
if line[0] != '#':
if line.split()[2] == 'gene':
disable_gene = True
elif line.split()[2] == 'transcript':
disable_trans = True
l += 1
if (disable_gene and disable_trans) or l == 100: break
ut.err_format_time(
'restore_gtf_db',
'Creating GTF databases for {} ...'.format(gtf_fn))
gtf_db = gu.create_db(gtf_fn,
gtf_db_fn,
disable_infer_genes=disable_gene,
disable_infer_transcripts=disable_trans)
ut.err_format_time(
'restore_gtf_db',
'Creating GTF databases for {} done!'.format(gtf_fn))
except:
ut.err_format_time(
'restore_gtf_db',
'Error in parsing {}\nCheck if annotation file format is correct'
.format(gtf_fn))
sys.exit(IOError)
else:
try:
ut.err_format_time(
'restore_gtf_db',
'Retrieving gff database for {} ...'.format(gtf_fn))
gtf_db = gu.FeatureDB(gtf_db_fn)
ut.err_format_time(
'restore_gtf_db',
'Retrieving gff database for {} done!'.format(gtf_fn))
except:
ut.err_format_time(
'restore_gtf_db',
'Error in parsing {}\nTry to remove this db file and re-run'
.format(gtf_db_fn))
sys.exit(IOError)
return gtf_db