def make_consensus(data_folder, adaID, fragment, n_iter, qual_min=20, VERBOSE=0,
coverage_min=10, summary=True):
'''Make consensus sequence from the mapped reads'''
if VERBOSE:
print 'Build consensus: '+adaID+' '+fragment+' iteration '+str(n_iter)
# Read reference
reffilename = get_reference_filename(data_folder, adaID, fragment, n_iter)
refseq = SeqIO.read(reffilename, 'fasta')
ref = np.array(refseq)
# Open BAM file
bamfilename = get_mapped_filename(data_folder, adaID, fragment, n_iter)
if not os.path.isfile(bamfilename):
convert_sam_to_bam(bamfilename)
(counts, inserts) = get_allele_counts_insertions_from_file_unfiltered(bamfilename,\
len(refseq), qual_min=qual_min,
match_len_min=match_len_min)
consensus_final = build_consensus(counts, inserts,
coverage_min=coverage_min,
VERBOSE=VERBOSE)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('Consensus built for iteration '+str(n_iter))
f.write('\n')
return refseq, consensus_final
def make_index_and_hash(data_folder, adaID, fragment, n_iter, VERBOSE=0,
summary=True):
'''Make index and hash files for reference or consensus'''
if VERBOSE:
print 'Build stampy hashes: '+adaID+' '+fragment+' iteration '+str(n_iter)
# 1. Make genome index file for 6 fragments (chromosomes)
call_list = [stampy_bin,
'--species="HIV adaID '+adaID+' fragment '+fragment+'"',
'--overwrite',
'-G', get_index_file(data_folder, adaID, fragment, n_iter, ext=False),
get_reference_filename(data_folder, adaID, fragment, n_iter),
]
if VERBOSE >= 3:
print ' '.join(call_list)
sp.call(call_list)
# 2. Build a hash file for 6 fragments
call_list = [stampy_bin,
'--overwrite',
'-g', get_index_file(data_folder, adaID, fragment, n_iter, ext=False),
'-H', get_hash_file(data_folder, adaID, fragment, n_iter, ext=False),
]
if VERBOSE >= 3:
print ' '.join(call_list)
sp.call(call_list)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('Made index and hash files for iteration '+str(n_iter))
f.write('\n')
def map_stampy(data_folder, adaID, fragment, n_iter, VERBOSE=0, summary=True):
'''Map using stampy'''
if VERBOSE:
print 'Map via stampy: '+adaID+' '+fragment+' iteration '+str(n_iter)
# Input and output files
input_filename = get_mapped_filename(data_folder, adaID, fragment,
n_iter - 1, type='bam')
output_filename = get_mapped_filename(data_folder, adaID, fragment,
n_iter, type='sam')
# Map
call_list = [stampy_bin,
'-g', get_index_file(data_folder, adaID, fragment,
n_iter, ext=False),
'-h', get_hash_file(data_folder, adaID, fragment,
n_iter, ext=False),
'-o', output_filename,
'--overwrite',
'--substitutionrate='+subsrate,
'--gapopen', stampy_gapopen,
'--gapextend', stampy_gapextend]
if stampy_sensitive:
call_list.append('--sensitive')
call_list = call_list + ['-M', input_filename]
call_list = map(str, call_list)
if VERBOSE >=2:
print ' '.join(call_list)
sp.call(call_list)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('Mapped completed for iteration '+str(n_iter))
f.write('\n')
def extract_reads_subsample(data_folder,
adaID,
fragment,
n_reads,
VERBOSE=0,
summary=True):
'''Extract a subsample of reads from the initial sample premapped'''
from hivwholeseq.utils.mapping import extract_mapped_reads_subsample
input_filename = get_divided_filename(data_folder,
adaID,
fragment,
type='bam')
output_filename = get_mapped_filename(data_folder,
adaID,
fragment,
n_iter=1,
type='bam')
n_written = extract_mapped_reads_subsample(input_filename,
output_filename,
n_reads,
VERBOSE=VERBOSE)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('\n')
f.write('Subsample of reads copied: ' + str(n_written))
f.write('\n')
def extract_reads_subsample(data_folder, adaID, fragment, n_reads, VERBOSE=0,
summary=True):
'''Extract a subsample of reads from the initial sample premapped'''
from hivwholeseq.utils.mapping import extract_mapped_reads_subsample
input_filename = get_divided_filename(data_folder, adaID, fragment, type='bam')
output_filename = get_mapped_filename(data_folder, adaID, fragment, n_iter=1, type='bam')
n_written = extract_mapped_reads_subsample(input_filename, output_filename,
n_reads, VERBOSE=VERBOSE)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('\n')
f.write('Subsample of reads copied: '+str(n_written))
f.write('\n')
def make_index_and_hash(data_folder,
adaID,
fragment,
n_iter,
VERBOSE=0,
summary=True):
'''Make index and hash files for reference or consensus'''
if VERBOSE:
print 'Build stampy hashes: ' + adaID + ' ' + fragment + ' iteration ' + str(
n_iter)
# 1. Make genome index file for 6 fragments (chromosomes)
call_list = [
stampy_bin,
'--species="HIV adaID ' + adaID + ' fragment ' + fragment + '"',
'--overwrite',
'-G',
get_index_file(data_folder, adaID, fragment, n_iter, ext=False),
get_reference_filename(data_folder, adaID, fragment, n_iter),
]
if VERBOSE >= 3:
print ' '.join(call_list)
sp.call(call_list)
# 2. Build a hash file for 6 fragments
call_list = [
stampy_bin,
'--overwrite',
'-g',
get_index_file(data_folder, adaID, fragment, n_iter, ext=False),
'-H',
get_hash_file(data_folder, adaID, fragment, n_iter, ext=False),
]
if VERBOSE >= 3:
print ' '.join(call_list)
sp.call(call_list)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('Made index and hash files for iteration ' + str(n_iter))
f.write('\n')
def make_consensus(data_folder,
adaID,
fragment,
n_iter,
qual_min=20,
VERBOSE=0,
coverage_min=10,
summary=True):
'''Make consensus sequence from the mapped reads'''
if VERBOSE:
print 'Build consensus: ' + adaID + ' ' + fragment + ' iteration ' + str(
n_iter)
# Read reference
reffilename = get_reference_filename(data_folder, adaID, fragment, n_iter)
refseq = SeqIO.read(reffilename, 'fasta')
ref = np.array(refseq)
# Open BAM file
bamfilename = get_mapped_filename(data_folder, adaID, fragment, n_iter)
if not os.path.isfile(bamfilename):
convert_sam_to_bam(bamfilename)
(counts, inserts) = get_allele_counts_insertions_from_file_unfiltered(bamfilename,\
len(refseq), qual_min=qual_min,
match_len_min=match_len_min)
consensus_final = build_consensus(counts,
inserts,
coverage_min=coverage_min,
VERBOSE=VERBOSE)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('Consensus built for iteration ' + str(n_iter))
f.write('\n')
return refseq, consensus_final
def map_stampy(data_folder, adaID, fragment, n_iter, VERBOSE=0, summary=True):
'''Map using stampy'''
if VERBOSE:
print 'Map via stampy: ' + adaID + ' ' + fragment + ' iteration ' + str(
n_iter)
# Input and output files
input_filename = get_mapped_filename(data_folder,
adaID,
fragment,
n_iter - 1,
type='bam')
output_filename = get_mapped_filename(data_folder,
adaID,
fragment,
n_iter,
type='sam')
# Map
call_list = [
stampy_bin, '-g',
get_index_file(data_folder, adaID, fragment, n_iter, ext=False), '-h',
get_hash_file(data_folder, adaID, fragment, n_iter, ext=False), '-o',
output_filename, '--overwrite', '--substitutionrate=' + subsrate,
'--gapopen', stampy_gapopen, '--gapextend', stampy_gapextend
]
if stampy_sensitive:
call_list.append('--sensitive')
call_list = call_list + ['-M', input_filename]
call_list = map(str, call_list)
if VERBOSE >= 2:
print ' '.join(call_list)
sp.call(call_list)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('Mapped completed for iteration ' + str(n_iter))
f.write('\n')
refseq, consensus = make_consensus(data_folder, adaID, fragment,
n_iter,
VERBOSE=VERBOSE, summary=summary)
write_consensus_intermediate(data_folder, adaID, fragment, n_iter, consensus)
match = check_new_old_consensi(refseq, consensus)
# Start a new round if not converged
if (not match) and (n_iter < iterations_max):
n_iter += 1
if VERBOSE:
print 'Starting again for iteration '+str(n_iter)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment), 'a') as f:
f.write('\n')
f.write('Starting new iteration '+str(n_iter))
f.write('\n')
# or terminate
else:
write_consensus_final(seq_run, adaID, fragment, consensus)
if VERBOSE:
if match:
print 'Consensus converged at iteration '+str(n_iter)+\
': adaID', adaID, fragment
else:
print 'Maximal number of iterations reached: adaID', \
adaID, fragment
VERBOSE=VERBOSE,
summary=summary)
write_consensus_intermediate(data_folder, adaID, fragment,
n_iter, consensus)
match = check_new_old_consensi(refseq, consensus)
# Start a new round if not converged
if (not match) and (n_iter < iterations_max):
n_iter += 1
if VERBOSE:
print 'Starting again for iteration ' + str(n_iter)
if summary:
with open(get_summary_fn(data_folder, adaID, fragment),
'a') as f:
f.write('\n')
f.write('Starting new iteration ' + str(n_iter))
f.write('\n')
# or terminate
else:
write_consensus_final(seq_run, adaID, fragment, consensus)
if VERBOSE:
if match:
print 'Consensus converged at iteration '+str(n_iter)+\
': adaID', adaID, fragment
else:
print 'Maximal number of iterations reached: adaID', \
adaID, fragment