Sex-aware Hardy-Weinberg screening challenges a standard GWAS quality step
Hardy-Weinberg equilibrium has long served as a routine quality-control filter in genome-wide association studies, but the framework was built for a simplified assumption: one homogeneous population, one shared genotype distribution, and little attention to sex-specific structure. This new analysis pushes back on that convention by asking a basic but clinically relevant genomics question: what happens when the same test is applied more carefully, with sex separated and with modern reference-aligned sequence data?
Using T2Tv2-aligned high-coverage whole-genome sequencing from 2,490 people in the 1000 Genomes Project, the investigators examined genome-wide deviations from Hardy-Weinberg equilibrium across five super-populations. Their focus was narrow and deliberate. Only bi-allelic single-nucleotide polymorphisms with non-missing genotypes and minor allele frequency of at least 5% in both sexes were retained, limiting the analysis to markers where the signal would be less likely to reflect sparse data or sequencing noise alone. The decision to work within a sex-aware framework matters because autosomal loci can show sex-linked distortion when analyses ignore biological differences or when sex chromosome sequence contaminates apparently autosomal calls.
One sentence captures the central concern: a familiar QC filter can behave differently once sex is no longer hidden inside a pooled sample.
Why the old filter can miss biologically meaningful structure
In genome studies, Hardy-Weinberg testing is often used as a blunt instrument to flag markers that may be technically unreliable, badly genotyped, or affected by population substructure. That approach has practical value, but it also assumes that departures from equilibrium are mostly nuisance signals. The biology is messier. Sex-based selection at autosomal loci has already been suggested in cohorts with active recruitment, and the presence of sequence from sex chromosomes can distort calls at nearby autosomal SNPs. When those influences are averaged across sexes, the resulting test statistic may lose directionality and conceal whether one sex carries the excess heterozygosity or deficit of expected genotypes.
The study therefore used an allele-based framework designed to quantify both the magnitude and direction of Hardy-Weinberg disequilibrium. That is an important methodological shift. Instead of simply asking whether a locus departs from equilibrium, the analysis asks how it departs, and whether the pattern is consistent across sexes and populations. In genetic epidemiology, direction can be as informative as significance. It helps distinguish a true biological asymmetry from a technical artifact that happens to produce the same p value.
T2T-aligned reference data sharpen the signal
The use of telomere-to-telomere aligned whole-genome data is not a cosmetic upgrade. The T2T framework improves representation of complex genomic regions, reduces mapping ambiguity, and should in principle lower the risk that a locus appears abnormal simply because it was aligned against an incomplete or imperfect reference. For Hardy-Weinberg analyses, that matters because false disequilibrium can emerge when reads map poorly, when nearby repetitive structure confuses genotype calling, or when sex chromosome-derived sequence bleeds into autosomal assessments. A cleaner reference should help separate genuine disequilibrium from reference-driven error.
The analysis also incorporated a second-order omnibus meta-analysis combining HWD results across populations and sexes. That design acknowledges a tension common to large genomic datasets: stratification improves biological resolution, but it also fragments sample size. Meta-analytic synthesis restores power while preserving the sex-specific structure that pooled analyses erase. In practice, this creates a more demanding QC screen, one that can identify loci that are stable in one sex but distorted in another, or that behave differently across ancestral groups.
What this means for association studies and marker filtering
The implications are immediate for GWAS pipelines. If Hardy-Weinberg testing continues to be treated as a one-size-fits-all filter, then sex-dependent effects may be misclassified as genotyping errors or, just as problematic, technical artifacts may be allowed to pass because they are diluted in a combined-sex analysis. In a field where millions of variants are screened, even modest distortion at the marker level can cascade into downstream false positives, false negatives, or inconsistent replication between cohorts.
This is especially relevant in studies that include multiple ancestral backgrounds, because population structure and sex structure do not act independently. A variant may appear to violate equilibrium in a pooled sample simply because allele frequencies differ between groups, or because the sex distribution is uneven across recruitment strata. The analytical choice to test by sex and then meta-analyze across super-populations gives the investigator a way to map where the deviation arises rather than collapsing all sources of heterogeneity into a single statistic.
A more discriminating QC standard is likely to be less forgiving
This approach will probably make QC stricter, not easier. Variants that once looked acceptable in a pooled screen may now fail a sex-aware test, and markers previously discarded on the basis of a single global p value may prove stable when sex is handled properly. Either outcome improves interpretability. The result is a more anatomically precise view of the genome, where each marker is judged in the context of the population and sex architecture in which it was measured.
The study also reinforces a broader principle in human genetics: reference quality and analytical design are inseparable. Better genome assemblies do not just improve variant discovery; they alter the statistical behavior of the tests used to validate those variants. In that sense, the move to T2T-aligned data is not merely a technical refinement. It changes the baseline from which equilibrium is judged.
A stricter framework still leaves unanswered questions
The abstract presents a strong rationale and a robust analytic design, but it does not yet provide the full catalog of loci affected, the distribution of findings across the five super-populations, or the extent to which sex-specific deviations mapped to known technical or biological mechanisms. Those results will determine how readily this sex-aware HWE framework can be adopted in routine GWAS pipelines.
What is already clear is narrower and more concrete: when Hardy-Weinberg equilibrium is assessed in high-quality whole-genome data, sex should no longer be treated as a nuisance variable. It is part of the genotype structure itself, and ignoring it risks flattening signals that matter for both quality control and biological inference.