Abstract Body (Do not enter title and authors here): Background:
Hypertrophic cardiomyopathy (HCM) is the most common inherited cardiomyopathy. As genomic sequencing becomes more widespread, early identification of at-risk individuals has improved. InterVar, which applies the 2015 ACMG guidelines, often classifies HCM-associated variants as likely pathogenic/pathogenic (LP/P) at frequencies exceeding true disease prevalence. Improved tools are needed to distinguish pathogenic from benign variants and to guide accurate population risk prediction.

Hypothesis:
HCM-associated variants classified as LP/P by our novel DiscoVari precision medicine tool demonstrate higher disease penetrance than those classified as LP/P by InterVar.

Methods:
Phenotypic and genotypic data from the UK Biobank (UKBB) and All of Us (AoU) cohorts were analyzed, focusing on missense variants in ClinGen-designated definitive HCM genes. DiscoVari performed signal-to-noise (S:N) analysis to define pathogenic “hotspots” by comparing variant frequencies in disease vs. population cohorts. Variants were reclassified per ACMG guidelines, using S:N hotspots to support PM1 evidence. Penetrance was defined as the proportion of individuals with clinical HCM among those with an HCM-associated variant. DiscoVari S:N and hotspot localization were also compared to variant classifications by Labcorp Genetics. Statistical significance was set at α = 0.05; Kruskal–Wallis tests were used for non-normal continuous data, and Chi-squared or Fisher’s exact tests for categorical data.

Results:
DiscoVari is key for variant interpretation as current ACMG criteria hinges on PM1 criteria in 26.5% (UKBB) and 21.7% (AoU) of cases. DiscoVari is superior in variant interpretation, reducing LP/P classifications among individuals without HCM (UKBB: 26.1% to 12%; AoU: 22% to 10.5%). DiscoVari LP/P variants were more strongly associated with HCM (UKBB: OR 2.75; AoU: OR 4.7) than InterVar LP/P variants (UKBB: OR 1.74; AoU: OR 4.07). Variants upgraded by DiscoVari had higher penetrance (OR 7.36) than downgraded variants, which showed no significant difference from the overall cohort. DiscoVari can aid in diagnostic variant interpretation as downgraded Labcorp diagnostic variants (n=578) had significantly lower median S:N than upgraded variants (n=313).

Conclusions:
Current in silico tools may misclassify low-penetrant variants in HCM genes as LP/P. DiscoVari improves interpretation by reducing over-classification of these variants while retaining those truly associated with disease.