def check_coverage(data_folder,
adaID,
fragment,
seq_run,
qual_min=35,
reference='HXB2',
maxreads=-1,
VERBOSE=0,
rescue=False,
minor_allele=False):
'''Check division into fragments: coverage, etc.'''
ref_fn = get_consensus_filename(data_folder, adaID, fragment)
refseq = SeqIO.read(ref_fn, 'fasta')
input_filename = get_mapped_filename(data_folder,
adaID,
fragment,
type='bam',
rescue=rescue)
counts, inserts = get_allele_counts_insertions_from_file_unfiltered(
input_filename, len(refseq), maxreads=maxreads, VERBOSE=VERBOSE)
# Plot results
title = ', '.join(
map(lambda x: ' '.join([x[0], str(x[1])]), [
['run', seq_run],
['adaID', adaID],
['fragment', fragment],
['maxreads', maxreads],
]))
plot_coverage(counts, suptitle=title, minor_allele=minor_allele)
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 get_allele_frequency_trajectories(pname,
samples,
fragment,
qual_min=30,
VERBOSE=0):
'''Scan the reads of all samples and write to a single file'''
if VERBOSE >= 1:
print 'Getting allele frequency trajectories:', pname, fragment
from hivwholeseq.patients.filenames import get_initial_reference_filename, \
get_mapped_to_initial_filename, get_allele_frequency_trajectories_filename, \
get_allele_count_trajectories_filename
from hivwholeseq.utils.one_site_statistics import get_allele_counts_insertions_from_file, \
get_allele_counts_insertions_from_file_unfiltered, \
filter_nus
refseq = SeqIO.read(get_initial_reference_filename(pname, fragment),
'fasta')
# Prepare output data structures
cos_traj = np.zeros((len(samples), len(alpha), len(refseq)), int)
nus_traj = np.zeros((len(samples), len(alpha), len(refseq)))
for it, sample in enumerate(samples):
if VERBOSE >= 2:
print pname, it, sample
input_filename = get_mapped_to_initial_filename(pname,
sample,
fragment,
type='bam')
(counts, inserts) = get_allele_counts_insertions_from_file_unfiltered(
input_filename, len(refseq), qual_min=qual_min, VERBOSE=VERBOSE)
# Take the total counts, blending in the read types
cou = counts.sum(axis=0)
cos_traj[it] = cou
# Take the filtered frequencies, blending in the read types
nu = filter_nus(counts)
nus_traj[it] = nu
#FIXME: test, etc.
return (cos_traj, nus_traj)
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 check_coverage(data_folder, adaID, fragment, seq_run, qual_min=35,
reference='HXB2', maxreads=-1, VERBOSE=0,
rescue=False,
minor_allele=False):
'''Check division into fragments: coverage, etc.'''
ref_fn = get_consensus_filename(data_folder, adaID, fragment)
refseq = SeqIO.read(ref_fn, 'fasta')
input_filename = get_mapped_filename(data_folder, adaID, fragment, type='bam',
rescue=rescue)
counts, inserts = get_allele_counts_insertions_from_file_unfiltered(input_filename,
len(refseq),
maxreads=maxreads,
VERBOSE=VERBOSE)
# Plot results
title=', '.join(map(lambda x: ' '.join([x[0], str(x[1])]),
[['run', seq_run],
['adaID', adaID],
['fragment', fragment],
['maxreads', maxreads],
]))
plot_coverage(counts, suptitle=title, minor_allele=minor_allele)
def check_premap(data_folder, adaID, fragments, seq_run, samplename,
qual_min=30, match_len_min=10,
maxreads=-1, VERBOSE=0,
savefig=True,
title=None):
'''Check premap to reference: coverage, etc.'''
refseq = SeqIO.read(get_reference_premap_filename(data_folder, adaID), 'fasta')
# FIXME: do this possibly better than parsing the description!
try:
fields = refseq.description.split()
refseq_start = int(fields[fields.index('(indices') - 3])
except ValueError:
refseq_start = 550
fragpos_filename = get_fragment_positions_filename(data_folder, adaID)
if os.path.isfile(fragpos_filename):
# Load the fragment positions, considering mixed fragments (e.g. F5a+b)
fragtmp = []
postmp = []
with open(fragpos_filename, 'r') as f:
f.readline() #HEADER
for line in f:
fields = line[:-1].split('\t')
fragtmp.append(fields[0])
if 'inner' not in fields[1]:
postmp.append([fields[1], fields[4]])
else:
start = int(fields[1].split(',')[1].split(': ')[1].rstrip('}'))
end = int(fields[4].split(',')[1].split(': ')[1].rstrip('}'))
postmp.append([start, end])
postmp = np.array(postmp, int)
# NOTE: In a lot of old files, it says F3o instead of F3ao
if 'F3o' in fragtmp:
fragtmp[fragtmp.index('F3o')] = 'F3ao'
elif 'F3i' in fragtmp:
fragtmp[fragtmp.index('F3i')] = 'F3ai'
frags_pos = np.array([postmp[fragtmp.index(fr)] for fr in fragments], int).T
else:
frags_pos = None
frags_pos_out = None
# Open BAM and scan reads
input_filename = get_premapped_filename(data_folder, adaID, type='bam')
if not os.path.isfile(input_filename):
if VERBOSE:
print 'Premapped BAM file not found'
return (None, None)
# Count reads if requested
n_reads = get_number_reads(input_filename)
if VERBOSE:
print 'N. of reads:', n_reads
# Get counts
counts, inserts = get_allele_counts_insertions_from_file_unfiltered(input_filename,
len(refseq),
qual_min=qual_min,
match_len_min=match_len_min,
maxreads=maxreads,
VERBOSE=VERBOSE)
# Plot results
if title is None:
title=', '.join(['run '+seq_run+' '+adaID,
'sample '+samplename,
'reads '+str(min(maxreads, n_reads))+'/'+str(n_reads),
])
plot_coverage(counts,
offset_x=refseq_start,
frags_pos=frags_pos,
frags_pos_out=frags_pos_out,
title=title)
if savefig:
from hivwholeseq.sequencing.adapter_info import foldername_adapter
plt.savefig(data_folder+foldername_adapter(adaID)+'figures/coverage_premapped_'+samplename+'.png')
return (counts, inserts)
def check_premap(data_folder,
adaID,
fragments,
seq_run,
samplename,
qual_min=30,
match_len_min=10,
maxreads=-1,
VERBOSE=0,
savefig=True,
title=None):
'''Check premap to reference: coverage, etc.'''
refseq = SeqIO.read(get_reference_premap_filename(data_folder, adaID),
'fasta')
# FIXME: do this possibly better than parsing the description!
try:
fields = refseq.description.split()
refseq_start = int(fields[fields.index('(indices') - 3])
except ValueError:
refseq_start = 550
fragpos_filename = get_fragment_positions_filename(data_folder, adaID)
if os.path.isfile(fragpos_filename):
# Load the fragment positions, considering mixed fragments (e.g. F5a+b)
fragtmp = []
postmp = []
with open(fragpos_filename, 'r') as f:
f.readline() #HEADER
for line in f:
fields = line[:-1].split('\t')
fragtmp.append(fields[0])
if 'inner' not in fields[1]:
postmp.append([fields[1], fields[4]])
else:
start = int(
fields[1].split(',')[1].split(': ')[1].rstrip('}'))
end = int(
fields[4].split(',')[1].split(': ')[1].rstrip('}'))
postmp.append([start, end])
postmp = np.array(postmp, int)
# NOTE: In a lot of old files, it says F3o instead of F3ao
if 'F3o' in fragtmp:
fragtmp[fragtmp.index('F3o')] = 'F3ao'
elif 'F3i' in fragtmp:
fragtmp[fragtmp.index('F3i')] = 'F3ai'
frags_pos = np.array([postmp[fragtmp.index(fr)] for fr in fragments],
int).T
else:
frags_pos = None
frags_pos_out = None
# Open BAM and scan reads
input_filename = get_premapped_filename(data_folder, adaID, type='bam')
if not os.path.isfile(input_filename):
if VERBOSE:
print 'Premapped BAM file not found'
return (None, None)
# Count reads if requested
n_reads = get_number_reads(input_filename)
if VERBOSE:
print 'N. of reads:', n_reads
# Get counts
counts, inserts = get_allele_counts_insertions_from_file_unfiltered(
input_filename,
len(refseq),
qual_min=qual_min,
match_len_min=match_len_min,
maxreads=maxreads,
VERBOSE=VERBOSE)
# Plot results
if title is None:
title = ', '.join([
'run ' + seq_run + ' ' + adaID,
'sample ' + samplename,
'reads ' + str(min(maxreads, n_reads)) + '/' + str(n_reads),
])
plot_coverage(counts,
offset_x=refseq_start,
frags_pos=frags_pos,
frags_pos_out=frags_pos_out,
title=title)
if savefig:
from hivwholeseq.sequencing.adapter_info import foldername_adapter
plt.savefig(data_folder + foldername_adapter(adaID) +
'figures/coverage_premapped_' + samplename + '.png')
return (counts, inserts)
