Individuals who deviate from polygenic expectation are enriched for damaging variants in genes linked to rare disease

Polygenic scores (PGSs) stratify disease risk but often fail to capture individual variation. "Misaligned" individuals, whose observed phenotypes deviate from their genetically expected values based on PGS, provide a powerful model for identifying factors beyond common-variant effects, including additional genetic factors. Here, we apply misalignment classification and enrichment testing frameworks to seven continuous traits and three diseases, assessing whether misaligned individuals in the UK Biobank are enriched for rare (minor-allele frequency [MAF] <0.1%) damaging genetic variation. We identified significant enrichment of predicted loss-of-function (pLoF) variants in COPB2 and GORAB among individuals with lower-than-expected bone mineral density. Regarding stature, shorter-than-expected individuals were enriched for pLoF variants in ACAN and IGF1, while taller-than-expected individuals showed enrichment for damaging missense variants in FBN1. Transitioning from validation to discovery, we performed an exome-wide scan for genes associated with misalignment and identified 74 significant genes, including KANK1, a gene which may have a protective role against primary ovarian insufficiency, and ACSL6, a lipid metabolism gene where damaging missense variation was associated with lower-than-expected BMI. For diseases, results supported a liability threshold model involving counteracting common and rare variant effects. Diagnosed type 2 diabetes mellitus patients with rare pathogenic variants in HNF1A and HNF4A possessed significantly lower polygenic risk than those without. Conversely, coronary artery disease controls harboring protective ANGPTL3 variants had nominally higher polygenic risk. This study highlights the power of misalignment-based analyses in complex continuous phenotypes and disease with the potential to validate known genetic contributors to traits and identify previously unassociated genes.