def cluster(
records,
threshold=0.95,
unknown='N',
key='gi',
aln_program='mafft',
aln_executable='mafft',
aln_options='--auto --reorder --adjustdirection',
seeds=None,
seed_coverage=0.5,
query_coverage=0.9,
free_gaps=True):
from krpy import krother
from krpy import krcl
results_dict = dict()
consumed_ids = list()
seed_ids = list()
record_count = len(records)
records = sorted(records, key=lambda x: len(x.seq), reverse=True)
records_seeds = records
if seeds:
records_seeds = seeds
for seed_rec in records_seeds:
key_value = None
if key == 'accession':
key_value = seed_rec.id
elif key == 'gi':
key_value = seed_rec.annotations['gi']
elif key == 'description':
key_value = seed_rec.description
else:
key_value = seed_rec.id
s_id = key_value
seed_ids.append(s_id)
for a_rec in records_seeds:
# print('a_rec', a_rec)
key_value = None
if key == 'accession':
key_value = a_rec.id
elif key == 'gi':
key_value = a_rec.annotations['gi']
elif key == 'description':
key_value = a_rec.description
else:
key_value = a_rec.id
a_id = key_value
if not seeds:
if a_id in consumed_ids:
continue
results_dict[a_id] = list()
if a_id not in consumed_ids:
results_dict[a_id].append(['+', a_id, '1.0'])
consumed_ids.append(a_id)
for i, b_rec in enumerate(records):
krcl.print_progress(
current=len(consumed_ids), total=record_count, length=0,
prefix=krother.timestamp() + ' ',
postfix=' records clustered. Checking ' + str(i) + '/' + str(record_count) + '.',
show_bar=False)
# print('b_rec', b_rec)
key_value = None
if key == 'accession':
key_value = b_rec.id
elif key == 'gi':
key_value = b_rec.annotations['gi']
elif key == 'description':
key_value = b_rec.description
else:
key_value = b_rec.id
b_id = key_value
if a_id == b_id:
continue
if b_id in consumed_ids:
continue
aln = align(
records=[a_rec, b_rec],
program=aln_program,
options=aln_options,
program_executable=aln_executable)
# print(aln)
pw_cov = pairwise_coverage(pairwise_alignment=aln)
# print(pw_cov)
a_cov = pw_cov[0]
b_cov = pw_cov[1]
# if a_rec.id in pw_cov.keys():
# a_cov = pw_cov[a_rec.id]
# else:
# a_cov = pw_cov['_R_' + a_rec.id]
# if b_rec.id in pw_cov.keys():
# b_cov = pw_cov[b_rec.id]
# else:
# b_cov = pw_cov['_R_' + b_rec.id]
direction = '+'
for a in aln:
# This will only work with MAFFT!
if a.id.startswith('_R_'):
direction = '-'
break
# score = pairwise_identity(
# alignment=aln,
# unknown_letters=set(['N']),
# unknown_id=0.0,
# free_unknowns=True,
# gap_id=0.0,
# free_gaps=True,
# end_gap_id=0.0,
# free_end_gaps=True)
score = identity(
alignment=aln,
unknown_letters=set(['N']),
free_unknowns=True,
free_gaps=free_gaps,
free_end_gaps=True)
if (score >= threshold) and (a_cov >= seed_coverage) and (b_cov >= query_coverage):
results_dict[a_id].append([direction, b_id, score])
consumed_ids.append(b_id)
krcl.clear_line()
# print(a_id, ':', b_id, '=', score, '|', a_cov, b_cov)
# Report unclustered ids
results_dict['unclustered'] = list()
for rec in records:
key_value = None
if key == 'accession':
key_value = rec.id
elif key == 'gi':
key_value = rec.annotations['gi']
elif key == 'description':
key_value = rec.description
else:
key_value = rec.id
rec_id = key_value
if rec_id not in consumed_ids:
results_dict['unclustered'].append(['.', rec_id, '0.0'])
return results_dict
authority_file = args.authority_file
if args.log_dir:
log_dir = args.log_dir
records = krbioio.read_sequence_file(input_file, 'gb', ret_type='list')
ps = os.path.sep
tax_log_handle = krseqsearch.__tax_log_open(log_dir, ps)
tax_log_html_handle = krseqsearch.__tax_log_html_open(log_dir, ps)
#########
krcl.hide_cursor()
for i, record in enumerate(records):
krcl.print_progress(i, len(records), 50, '')
name = krseqsearch.check_organism_name(
record,
ncbi_names_table,
synonymy_table,
authority_file,
hacks,
hacks_data_location,
unresolvable_taxonomy_list,
keeplist_taxonomy_list,
taxa_mappings_list,
tax_log_handle,
tax_log_html_handle)
# tn = name[0]
