def haplotype_caller(core_req=16,
mem_req=12 * 1024,
in_bams=find('bam$', n='>0'),
in_bais=find('bai$', n='>0'),
in_target_bed=find('target.bed'),
out_vcf=out_dir('raw_variants.g.vcf')):
in_bams = bam_list_to_inputs(in_bams)
intervals = arg('--intervals', in_target_bed)
return r"""
{gatk} \
-T HaplotypeCaller \
-R {s[ref][reference_fasta]} \
-D {s[ref][dbsnp_vcf]} \
-nct {core_req} \
--emitRefConfidence GVCF \
-stand_call_conf 30 \
-stand_emit_conf 10 \
-I {in_bams} \ -o {out_vcf} \
{intervals} \
-A Coverage \
-A GCContent \
-A AlleleBalanceBySample \
-A AlleleBalance \
-A MappingQualityRankSumTest \
-A InbreedingCoeff \
-A FisherStrand \
-A QualByDepth
""".format(s=s, gatk=gatk(mem_req), **locals())
def indel_realigner(core_req=4, # proxy for mem_req until i test mem_req out
mem_req=8 * 1024,
contig=None,
in_bams=find('bam$', n='>0'),
in_bais=find('bai$', n='>0'),
in_sites=find('denovo_realign_targets.bed'),
out_bam=out_dir('realigned.bam'),
out_bai=out_dir('realigned.bai')):
in_bams = bam_list_to_inputs(in_bams)
if s['ref']['version'] == 'b37':
in_knowns = s['ref']['1kg_indel_vcf'], s['ref']['mills_vcf']
elif s['ref']['version'] == 'hg38':
in_knowns = [s['ref']['mills_and_1kg_indel_vcf']]
return r"""
# IR does not support parallelization
{gatk} \
-T IndelRealigner \
-R {s[ref][reference_fasta]} \
-I {in_bams} \
-o {out_bam} \
-targetIntervals {in_sites} \
{knowns} \
-model USE_READS \
--filter_bases_not_stored \
{intervals}
{s[opt][samtools]} index {out_bam}
""".format(s=s,
intervals=arg('--intervals', contig),
gatk=gatk(mem_req),
knowns=' '.join('-known %s' % p for p in in_knowns),
**locals())
def realigner_target_creator(core_req=8,
mem_req=8 * 1024,
in_target_bed=find('target.bed'),
in_bams=find('bam$', n='>0'),
in_bais=find('bai$', n='>0'),
out_bams=forward('in_bams'),
out_bais=forward('in_bais'),
out_sites=out_dir('denovo_realign_targets.bed')):
in_bams = bam_list_to_inputs(in_bams)
if s['ref']['version'] == 'b37':
in_knowns = s['ref']['1kg_indel_vcf'], s['ref']['mills_vcf']
elif s['ref']['version'] == 'hg38':
in_knowns = [s['ref']['mills_and_1kg_indel_vcf']]
# TODO should we pad intervals? might be indels on perimeter that need realigner. Not too worried because we're using HaplotypeCaller, though.
return r"""
#could add more knowns from ESP and other seq projects...
{gatk} \
-T RealignerTargetCreator \
-R {s[ref][reference_fasta]} \
-I {in_bams} \
-o {out_sites} \
{knowns} \
-nt {core_req} \
{args}
""".format(s=s, gatk=gatk(mem_req),
args=arg('--intervals', in_target_bed),
knowns=' '.join('-known %s' % p for p in in_knowns),
**locals())
def haplotype_caller(core_req=16,
mem_req=29 * 1024,
in_bams=find('bam$', n='>0'),
in_bais=find('bai$', n='>0'),
in_target_bed=find('target.bed'),
out_vcf=out_dir('raw_variants.g.vcf')):
in_bams = bam_list_to_inputs(in_bams)
intervals = arg('--intervals', in_target_bed)
return r"""
{gatk} \
-T HaplotypeCaller \
-R {s[ref][reference_fasta]} \
-D {s[ref][dbsnp_vcf]} \
-nct {core_req} \
--emitRefConfidence GVCF \
-I {in_bams} \
-o {out_vcf} \
{intervals} \
-A Coverage \
-A GCContent \
-A AlleleBalanceBySample \
-A AlleleBalance \
-A MappingQualityRankSumTest \
-A InbreedingCoeff \
-A FisherStrand \
-A QualByDepth
""".format(s=s, gatk=gatk(mem_req), **locals())
def indel_realigner(core_req=4, # proxy for mem_req until i test mem_req out
mem_req=8 * 1024,
contig=None,
in_bams=find('bam$', n='>0'),
in_bais=find('bai$', n='>0'),
in_sites=find('denovo_realign_targets.bed'),
out_bam=out_dir('realigned.bam'),
out_bai=out_dir('realigned.bai')):
in_bams = bam_list_to_inputs(in_bams)
if s['ref']['version'] == 'b37':
in_knowns = s['ref']['1kg_indel_vcf'], s['ref']['mills_vcf']
elif s['ref']['version'] == 'hg38':
in_knowns = [s['ref']['mills_and_1kg_indel_vcf']]
return r"""
# IR does not support parallelization
{gatk} \
-T IndelRealigner \
-R {s[ref][reference_fasta]} \
-I {in_bams} \
-o {out_bam} \
-targetIntervals {in_sites} \
{knowns} \
-model USE_READS \
--filter_bases_not_stored \
{intervals}
{s[opt][samtools]} index {out_bam}
""".format(s=s,
intervals=arg('--intervals', contig),
gatk=gatk(mem_req),
knowns=' '.join('-known %s' % p for p in in_knowns),
**locals())
def realigner_target_creator(core_req=8,
mem_req=8 * 1024,
in_target_bed=find('target.bed'),
in_bams=find('bam$', n='>0'),
in_bais=find('bai$', n='>0'),
out_bams=forward('in_bams'),
out_bais=forward('in_bais'),
out_sites=out_dir('denovo_realign_targets.bed')):
in_bams = bam_list_to_inputs(in_bams)
if s['ref']['version'] == 'b37':
in_knowns = s['ref']['1kg_indel_vcf'], s['ref']['mills_vcf']
elif s['ref']['version'] == 'hg38':
in_knowns = [s['ref']['mills_and_1kg_indel_vcf']]
# TODO should we pad intervals? might be indels on perimeter that need realigner. Not too worried because we're using HaplotypeCaller, though.
return r"""
#could add more knowns from ESP and other seq projects...
{gatk} \
-T RealignerTargetCreator \
-R {s[ref][reference_fasta]} \
-I {in_bams} \
-o {out_sites} \
{knowns} \
-nt {core_req} \
{args}
""".format(s=s, gatk=gatk(mem_req),
args=arg('--intervals', in_target_bed),
knowns=' '.join('-known %s' % p for p in in_knowns),
**locals())
