def _convert_to_py(self, annotation):
if annotation:
return Interval._from_java(annotation)
else:
return annotation
# * In tier 1 regions.
#
# Definition of probably not somatic:
# DP > 10 AND
# (
# GT != het OR
# (
# binomTest(ad, dp, 0.5, "two.sided") >= alpha
# )
# )
#
# With alpha = 0.1 we expect a little over 10% of het sites will be falsely lost.
# For alpha = 0.1, the DP > 10 restriction corresponds to effective bounds on VAF
# of [0.2, 0.8].
for chrom in range(1, 23):
print 'Chromosome %d...' % chrom
(vds.filter_intervals(Interval.parse(
'%d' % chrom)).annotate_variants_table(
tier1_bed, root='va.tier1bed').filter_variants_expr(
'va.tier1bed == true && v.isAutosomal()',
keep=True).filter_samples_expr(
'"^Z" ~ s', keep=False).split_multi().filter_genotypes(
'''
g.dp > 10 &&
(
(!g.isHet()) || binomTest(g.ad[1], g.dp, 0.5, "two.sided") >= 0.1
)''',
keep=True).min_rep().export_plink(
'tmp/08b_genotypes_%d' % chrom))
def _convert_to_py(self, annotation):
if annotation:
return Interval._from_java(annotation)
else:
return annotation
#!../../software/pyhail.sh
import hail
from hail.representation import Interval
from hail.expr import TString, TBoolean, TFloat, TInt
hc = hail.HailContext(log = 'log/99_dreamlab2.log', tmp_dir = 'tmp/hail')
vds = hc.read('../MGRB.phase2.tier12.match.vqsr.minrep.vds')
# Chr22 only, rough variant quality filters
vds = (vds
.filter_intervals(Interval.parse('22'), keep=True)
.filter_variants_expr('va.filters.isEmpty()', keep=True)
.split_multi()
.variant_qc()
.filter_variants_expr('''
v.altAllele.isSNP &&
va.qc.callRate >= 0.99 &&
va.qc.dpMean >= 20 && va.qc.dpMean <= 60 &&
va.qc.dpStDev < 8 &&
va.filters.isEmpty() &&
va.qc.AF >= 0.05 && va.qc.AF <= 0.95''')
)
# Drop samples with poor metrics on these filtered variants.
vds = (vds
.sample_qc()
.filter_samples_expr('sa.qc.callRate >= 0.985')
)
#!/usr/bin/env python
import argparse as ap
import hail
from hail.representation import Interval
p = ap.ArgumentParser()
p.add_argument("--exomes-vds", help="Exomes dataset to be loaded, already split", required=True)
p.add_argument("--genomes-vds", help="Genomes dataset to be loaded, already split", required=True)
p.add_argument( "--exomes-vds-out", help="Exomes file to be written", required=True)
p.add_argument("--genomes-vds-out", help="Genomes file to be written", required=True)
p.add_argument("--interval", help="Interval to subset", required=True)
args = p.parse_args()
hc = hail.HailContext(log="/hail.log")
(hc.read(args.exomes_vds)
.filter_intervals(Interval.parse(args.interval))
.write(args.exomes_vds_out, overwrite=True))
(hc.read(args.genomes_vds)
.filter_intervals(Interval.parse(args.interval))
.write(args.genomes_vds_out, overwrite=True))
#!./bin/pyhail.sh
import pyspark
import hail
from hail.representation import Interval
from hail import KeyTable
import os.path
hc = hail.HailContext(log='log/07_plink_export_45andup_subpops.log',
tmp_dir='tmp/hail')
interval_list = [Interval.parse('%d' % chrom) for chrom in range(1, 22)]
tier1_bed = KeyTable.import_bed(
'../../locus-annotations/source_data/HG001_GRCh37_GIAB_highconf_CG-IllFB-IllGATKHC-Ion-10X-SOLID_CHROM1-X_v.3.3.2_highconf_nosomaticdel.bed'
)
(hc.read(
'../MGRB.phase2.SNPtier12.match.vqsr.minrep.locusannot.WGStier12.unrelated.vds'
).filter_intervals(interval_list, keep=True).annotate_variants_table(
tier1_bed, root='va.tier1bed').filter_variants_expr(
'va.tier1bed == true', keep=True).annotate_variants_expr('va = {}').
filter_samples_expr('"^B" ~ s').split_multi().min_rep().write(
'tmp/MGRB.phase2.SNPtier12.match.vqsr.minrep.locusannot.WGStier12.unrelated.45andUp.GiaB_HCR.noannot.split.minrep.vds'
))
sample_lists = [
'../45andup_followup_qcpass_anycancer_mf.sample_list',
'../45andup_followup_qcpass_breastcancer_f.sample_list',
'../45andup_followup_qcpass_breastcancer_mf.sample_list',
'../45andup_followup_qcpass_colorectalcancer_mf.sample_list',
'../45andup_followup_qcpass_melanomacancer_mf.sample_list',
'../45andup_followup_qcpass_nocancer_f.sample_list',