def analyze_reads(fqs, paired, reads_file, keep_files):
'''Analyzes read length for single-end sampled, required by Kallisto.'''
awk = "| awk '{if(NR%4==2) print length($1)}'"
log.info('[alignment] Analyzing read length')
if paired:
fq1, fq2 = fqs
command = ['zcat <', fq1, awk, '>', reads_file]
run_command(command)
command = ['zcat <', fq2, awk, '>>', reads_file]
run_command(command)
else:
fq = fqs[0]
command = ['zcat <', fq, awk, '>', reads_file]
run_command(command)
read_lengths = np.genfromtxt(reads_file)
num = len(read_lengths)
avg = round(np.mean(read_lengths), 4)
std = round(np.std(read_lengths), 4)
remove_files([reads_file], keep_files)
return num, avg, std
def process_counts(count_file, eq_file, gene_list, allele_idx, allele_lengths,
keep_files):
'''Processes pseudoalignment output, returning compatibility classes.'''
log.info('[alignment] processing pseudoalignment')
# Process count information
counts = dict()
with open(count_file, 'r') as file:
for line in file.read().splitlines():
eq, count = line.split('\t')
counts[eq] = float(count)
# Process compatibility classes
eqs = dict()
with open(eq_file, 'r') as file:
for line in file.read().splitlines():
eq, indices = line.split('\t')
eqs[eq] = indices.split(',')
# Set up compatibility class index
eq_idx = defaultdict(list)
count_unique = 0
count_multi = 0
class_unique = 0
class_multi = 0
for eq, indices in eqs.items():
if [idx for idx in indices if not allele_idx[idx]]:
continue
genes = list({
get_gene(allele)
for idx in indices for allele in allele_idx[idx]
})
count = counts[eq]
if len(genes) == 1 and counts[eq] > 0:
gene = genes[0]
eq_idx[gene].append((indices, count))
count_unique += count
class_unique += 1
else:
count_multi += count
class_multi += 1
# Alleles mapping to their respective compatibility classes
allele_eq = defaultdict(set)
for eqs in eq_idx.values():
for eq, (indices, _) in enumerate(eqs):
for idx in indices:
allele_eq[idx].add(eq)
remove_files([count_file, eq_file], keep_files)
align_stats = [count_unique, count_multi, class_unique, class_multi]
return eq_idx, allele_eq, align_stats
def pseudoalign(fqs,
sample,
paired,
reference,
outdir,
temp,
threads,
keep_files,
partial=False):
'''Calls Kallisto to pseudoalign reads.'''
file_list = []
# Get read length stats
reads_file = ''.join([temp, sample, '.reads.txt'])
num, avg, std = analyze_reads(fqs, paired, reads_file, keep_files)
# Kallisto fails if std used for single-end is 0
if std == 0: std = .00001
temp2 = check_path(''.join([temp, sample]))
command = ['kallisto pseudo -i', reference, '-t', threads, '-o', temp2]
if paired:
command.extend([fqs[0], fqs[1]])
else:
fq = fqs[0]
command.extend(['--single -l', str(avg), '-s', str(std), fq])
run_command(command, '[alignment] Pseudoaligning with Kallisto: ')
# Move and rename Kallisto output
file_in = ''.join([temp2, 'pseudoalignments.tsv'])
count_file = ''.join([temp, sample, '.counts.tsv'])
file_list.append(file_in)
run_command(['mv', file_in, count_file])
file_in = ''.join([temp2, 'pseudoalignments.ec'])
eq_file = ''.join([temp, sample, '.eq.tsv'])
file_list.append(file_in)
run_command(['mv', file_in, eq_file])
run_command(['rm -rf', temp2])
remove_files(file_list, keep_files)
return count_file, eq_file, num, avg, std
def extract_reads(bam, outdir, paired, unmapped, alts,
temp, threads, keep_files):
'''Extracts reads from chromosome 6 and alts/decoys if applicable.'''
log.info(f'[extract] Extracting reads from {bam}')
file_list = []
sample = os.path.splitext(os.path.basename(bam))[0]
# Index bam
index_bam(bam)
hla_filtered = ''.join([temp, sample, '.hla.sam'])
file_list.append(hla_filtered)
hla_filtered_bam = ''.join([temp, sample, '.hla.bam'])
file_list.append(hla_filtered_bam)
# Get bam header to check for chromosome nomenclature
output = run_command(['samtools', 'view', '-@'+threads, '-H', bam])
header = output.stdout.decode('utf-8')
if 'SN:chr' in header:
chrom = 'chr6'
else:
chrom = '6'
# Extract BAM header
message = '[extract] Extracting chromosome 6: '
command = ['samtools', 'view', '-H', '-@'+threads]
command.extend([bam, '-o', hla_filtered])
run_command(command, message)
# Extracted reads mapped to chromosome 6
message = '[extract] Extracting chromosome 6: '
command = ['samtools', 'view', '-@'+threads]
if paired: command.append('-f 2')
else: command.append('-F 4')
command.extend([bam, chrom, '>>', hla_filtered])
run_command(command, message)
# Extract unmapped reads
if unmapped:
message = '[extract] Extracting chromosome 6: '
command = ['samtools', 'view', '-@'+threads]
if paired: command.append('-f 12')
else: command.append('-f 4')
command.extend([bam, chrom, '>>', hla_filtered])
run_command(command, message)
# Check for alts in header and extract reads if present
for alt in alts:
if alt in header:
command = ['samtools', 'view', '-@'+threads]
if paired: command.append('-f 2')
else: command.append('-F 4')
command.extend([bam, alt+':', '>>', hla_filtered])
run_command(command)
# Convert SAM to BAM
message = '[extract] Converting SAM to BAM: '
command = ['samtools', 'view', '-Sb', '-@'+threads,
hla_filtered, '>', hla_filtered_bam]
run_command(command, message)
# Sort BAM
hla_sorted = ''.join([temp, sample, '.hla.sorted.bam'])
file_list.append(hla_sorted)
message = '[extract] Sorting bam: '
command = ['samtools', 'sort', '-n', '-@'+threads,
hla_filtered_bam, '-o', hla_sorted]
run_command(command, message)
# Convert BAM to FASTQ and compress
message = '[extract] Converting bam to fastq: '
command = ['bedtools', 'bamtofastq', '-i', hla_sorted]
if paired:
fq1 = ''.join([outdir, sample, '.extracted.1.fq'])
fq2 = ''.join([outdir, sample, '.extracted.2.fq'])
command.extend(['-fq', fq1, '-fq2', fq2])
run_command(command, message)
run_command(['pigz', '-f', '-p', threads, '-S', '.gz', fq1])
run_command(['pigz', '-f', '-p', threads, '-S', '.gz', fq2])
else:
fq = ''.join([outdir, sample, '.extracted.fq'])
command.extend(['-fq', fq])
run_command(command, message)
run_command(['pigz', '-f', '-p', threads, '-S', '.gz', fq])
remove_files(file_list, keep_files)