def age_hists_expr(
adj_expr: hl.expr.BooleanExpression,
gt_expr: hl.expr.CallExpression,
age_expr: hl.expr.NumericExpression,
lowest_boundary: int = 30,
highest_boundary: int = 80,
n_bins: int = 10,
) -> hl.expr.StructExpression:
"""
Returns a StructExpression with the age histograms for hets and homs.
:param adj_expr: Entry expression containing whether a genotype is high quality (adj) or not
:param gt_expr: Entry expression containing the genotype
:param age_expr: Col expression containing the sample's age
:param lowest_boundary: Lowest bin boundary (any younger sample will be binned in n_smaller)
:param highest_boundary: Highest bin boundary (any older sample will be binned in n_larger)
:param n_bins: Total number of bins
:return: A struct with `age_hist_het` and `age_hist_hom`
"""
return hl.struct(
age_hist_het=hl.agg.filter(
adj_expr & gt_expr.is_het(),
hl.agg.hist(age_expr, lowest_boundary, highest_boundary, n_bins),
),
age_hist_hom=hl.agg.filter(
adj_expr & gt_expr.is_hom_var(),
hl.agg.hist(age_expr, lowest_boundary, highest_boundary, n_bins),
),
)
def _is_dnm(
proband_gt: hl.expr.CallExpression,
father_gt: hl.expr.CallExpression,
mother_gt: hl.expr.CallExpression,
locus: hl.expr.LocusExpression,
proband_is_female: Optional[hl.expr.BooleanExpression],
) -> hl.expr.BooleanExpression:
"""
Helper method to get whether a given genotype combination is a DNM at a given locus with a given proband sex.
"""
if proband_is_female is None:
logger.warning(
"Since no proband sex expression was given to generate_trio_stats_expr, only DNMs in autosomes will be counted."
)
return hl.or_missing(
locus.in_autosome(),
proband_gt.is_het() & father_gt.is_hom_ref()
& mother_gt.is_hom_ref(),
)
return hl.cond(
locus.in_autosome_or_par() |
(proband_is_female & locus.in_x_nonpar()),
proband_gt.is_het() & father_gt.is_hom_ref()
& mother_gt.is_hom_ref(),
hl.or_missing(~proband_is_female,
proband_gt.is_hom_var() & father_gt.is_hom_ref()),
)
