Evaluating the Role of Hypomorphic Variants in Cholestasis via Network Analysis Techniques

Hypomorphic genetic variants, which partially reduce gene function, may contribute to phenotypic variability beyond strongly deleterious mutations typically implicated in Mendelian diseases. We introduce a network-based statistical framework for assessing the structural role of hypomorphic variation in rare diseases, and its implementation as an R package, HyVarGeNet, supporting end-to-end analysis from annotated VCF files to patient-specific gene interaction networks. The framework was applied to intrahepatic cholestasis in a cohort comprising clinically affected individuals, parents, and healthy controls. Gene-level hypomorphic scores were integrated into STRING-derived interaction networks and evaluated using Conditional Uniform Graph tests to assess homophily. Exponential Random Graph Models were subsequently fitted to quantify the contribution of hypomorphic similarity to edge formation, adjusting for structural network effects and gene-level covariates. Results indicate that genes with comparable hypomorphic burden are significantly more likely to interact than expected under degree-preserving null models, and that this association remains significant in multivariable network models, with heterogeneity across study groups.