Natural genetic diversity modulates structural stability changes caused by pathogenic variants in small proteins

Clinical genomics relies heavily on linear reference genomes such as GRCh38, often neglecting natural sequence diversity that can modify variant expressivity and drive diagnostic health disparities. Meanwhile, variant interpretation is moving toward direct biophysical evidence of disrupted protein function, yet how natural genetic backgrounds modulate variant-induced protein stability changes remains uncharacterized. Here, we present a physics-based computational and data processing framework to evaluate how global sequence diversity alters the predicted thermodynamic impact of pathogenic mutations. Focusing on small human proteins (≤ 100 residues), we reconstructed natural DNA haplotypes from 1000 Genomes Project data and introduced ClinVar pathogenic variants, yielding both natural and pathogenic protein haplotypes that were structurally modeled and energetically relaxed. Across 271 tested variants in 46 genes, structural stability changes (ΔΔG$) for 168 variants (62%) differed significantly depending on the underlying natural haplotype background (Kruskal-Wallis, Bonferroni-adjusted p < 0.01). These background-sensitive variants spanned 43 of the 46 genes (93%), suggesting that thermodynamic modulation by natural sequence context is a pervasive phenomenon across the human proteome. Accounting for haplotype context in protein stability profiling is therefore essential for equitable, ancestry-aware precision medicine.