Identification of incompletely penetrant variants and interallelic interactions in autosomal recessive disorders by a population-genetic approach

We aimed to identify incompletely penetrant (IP) variants and interallelic interactions in autosomal recessive disorders by a population-genetic approach. Genotype and clinical data were collected from 9038 patients of European origin with ASL, ATP7B, CAPN3, CFTR, CTNS, DHCR7, GAA, GALNS, GALT, IDUA, MUT, NPHS1, NPHS2, PAH, PKHD1, PMM2, or SLC26A4-related disorders. We calculated the relative allele frequency of each pathogenic variant (n = 1936) to the loss-of-function (LOF) variants of the corresponding gene in the patient ( $A{C}_{pt}^V/A{C}_{pt}^{LOF}$ ) and the general population ( ${\mathrm{AC}}_{\mathrm{gnomAD}}^V/{\mathrm{AC}}_{\mathrm{gnomAD}}^{\mathrm{LOF}}$ ) and estimated the penetrance of each variant by calculating their ratio: $\frac{(A{C}_{pt}^V/A{C}_{pt}^{LOF})}{(A{C}_{gnomAD}^V/A{C}_{gnomAD}^{LOF})}$ (V/LOF ratio). We classified all variants as null or hypomorphic based on the associated clinical phenotype. We found 25 variants, 29% of the frequent 85 variants, to be underrepresented in the patient population (V/LOF ratio <30% with p < 7.22 × 10^-5 ), including 22 novel ones in the ASL, CAPN3, CFTR, GAA, GALNS, PAH, and PKHD1 genes. In contrast to the completely penetrant variants (CP), the majority of the IP variants were hypomorphic (IP: 16/18, 88%; CP: 177/933, 19.0%; p = 5.12 × 10^-10 ). Among them, only the NPHS2 R229Q variant was subject to interallelic interactions. The proposed algorithm identifies frequent IP variants and estimates their penetrance and interallelic interactions in large patient cohorts.