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()),
)
def _get_copy_state(
locus: hl.expr.LocusExpression) -> hl.expr.Int32Expression:
"""
Helper method to go from LocusExpression to a copy-state int for indexing into the
trans_count_map.
"""
return (hl.case().when(locus.in_autosome_or_par(), auto_or_par).when(
locus.in_x_nonpar(), hemi_x).when(locus.in_y_nonpar(),
hemi_y).or_missing())
def faf_expr(
freq: hl.expr.ArrayExpression,
freq_meta: hl.expr.ArrayExpression,
locus: hl.expr.LocusExpression,
pops_to_exclude: Optional[Set[str]] = None,
faf_thresholds: List[float] = [0.95, 0.99],
) -> Tuple[hl.expr.ArrayExpression, List[Dict[str, str]]]:
"""
Calculates the filtering allele frequency (FAF) for each threshold specified in `faf_thresholds`.
See http://cardiodb.org/allelefrequencyapp/ for more information.
The FAF is computed for each of the following population stratification if found in `freq_meta`:
- All samples, with adj criteria
- For each population, with adj criteria
- For all sex/population on the non-PAR regions of sex chromosomes (will be missing on autosomes and PAR regions of sex chromosomes)
Each of the FAF entry is a struct with one entry per threshold specified in `faf_thresholds` of type float64.
This returns a tuple with two expressions:
1. An array of FAF expressions as described above
2. An array of dict containing the metadata for each of the array elements, in the same format as that produced by `annotate_freq`.
:param freq: ArrayExpression of call stats structs (typically generated by hl.agg.call_stats)
:param freq_meta: ArrayExpression of meta dictionaries corresponding to freq (typically generated using annotate_freq)
:param locus: locus
:param pops_to_exclude: Set of populations to exclude from faf calculation (typically bottlenecked or consanguineous populations)
:param faf_thresholds: List of FAF thresholds to compute
:return: (FAF expression, FAF metadata)
"""
_pops_to_exclude = (
hl.literal(pops_to_exclude) if pops_to_exclude is not None else {}
)
# pylint: disable=invalid-unary-operand-type
faf_freq_indices = hl.range(0, hl.len(freq_meta)).filter(
lambda i: (freq_meta[i].get("group") == "adj")
& (
(freq_meta[i].size() == 1)
| (
(hl.set(freq_meta[i].keys()) == {"pop", "group"})
& (~_pops_to_exclude.contains(freq_meta[i]["pop"]))
)
)
)
sex_faf_freq_indices = hl.range(0, hl.len(freq_meta)).filter(
lambda i: (freq_meta[i].get("group") == "adj")
& (freq_meta[i].contains("sex"))
& (
(freq_meta[i].size() == 2)
| (
(hl.set(freq_meta[i].keys()) == {"pop", "group", "sex"})
& (~_pops_to_exclude.contains(freq_meta[i]["pop"]))
)
)
)
faf_expr = faf_freq_indices.map(
lambda i: hl.struct(
**{
f"faf{str(threshold)[2:]}": hl.experimental.filtering_allele_frequency(
freq[i].AC, freq[i].AN, threshold
)
for threshold in faf_thresholds
}
)
)
faf_expr = faf_expr.extend(
sex_faf_freq_indices.map(
lambda i: hl.or_missing(
~locus.in_autosome_or_par(),
hl.struct(
**{
f"faf{str(threshold)[2:]}": hl.experimental.filtering_allele_frequency(
freq[i].AC, freq[i].AN, threshold
)
for threshold in faf_thresholds
}
),
)
)
)
faf_meta = faf_freq_indices.extend(sex_faf_freq_indices).map(lambda i: freq_meta[i])
return faf_expr, hl.eval(faf_meta)
def get_summary_counts_dict(
locus_expr: hl.expr.LocusExpression,
allele_expr: hl.expr.ArrayExpression,
lof_expr: hl.expr.StringExpression,
no_lof_flags_expr: hl.expr.BooleanExpression,
most_severe_csq_expr: hl.expr.StringExpression,
prefix_str: str = "",
) -> Dict[str, hl.expr.Int64Expression]:
"""
Return dictionary containing containing counts of multiple variant categories.
Categories are:
- Number of variants
- Number of indels
- Number of SNVs
- Number of LoF variants
- Number of LoF variants that pass LOFTEE
- Number of LoF variants that pass LOFTEE without any flgs
- Number of LoF variants annotated as 'other splice' (OS) by LOFTEE
- Number of LoF variants that fail LOFTEE
- Number of missense variants
- Number of synonymous variants
- Number of autosomal variants
- Number of allosomal variants
.. warning::
Assumes `allele_expr` contains only two variants (multi-allelics have been split).
:param locus_expr: LocusExpression.
:param allele_expr: ArrayExpression containing alleles.
:param lof_expr: StringExpression containing LOFTEE annotation.
:param no_lof_flags_expr: BooleanExpression indicating whether LoF variant has any flags.
:param most_severe_csq_expr: StringExpression containing most severe consequence annotation.
:param prefix_str: Desired prefix string for category names. Default is empty str.
:return: Dict of categories and counts per category.
"""
logger.warning(
"This function expects that multi-allelic variants have been split!")
return {
f"{prefix_str}num_variants":
hl.agg.count(),
f"{prefix_str}indels":
hl.agg.count_where(hl.is_indel(allele_expr[0], allele_expr[1])),
f"{prefix_str}snps":
hl.agg.count_where(hl.is_snp(allele_expr[0], allele_expr[1])),
f"{prefix_str}LOF":
hl.agg.count_where(hl.is_defined(lof_expr)),
f"{prefix_str}pass_loftee":
hl.agg.count_where(lof_expr == "HC"),
f"{prefix_str}pass_loftee_no_flag":
hl.agg.count_where((lof_expr == "HC") & (no_lof_flags_expr)),
f"{prefix_str}loftee_os":
hl.agg.count_where(lof_expr == "OS"),
f"{prefix_str}fail_loftee":
hl.agg.count_where(lof_expr == "LC"),
f"{prefix_str}num_missense":
hl.agg.count_where(most_severe_csq_expr == "missense_variant"),
f"{prefix_str}num_synonymous":
hl.agg.count_where(most_severe_csq_expr == "synonymous_variant"),
f"{prefix_str}num_autosomal_variants":
hl.agg.filter(locus_expr.in_autosome_or_par(), hl.agg.count()),
f"{prefix_str}num_allosomal_variants":
hl.agg.filter(locus_expr.in_x_nonpar() | locus_expr.in_y_nonpar(),
hl.agg.count()),
}
