Towards improved missense variant prioritization through automated structure-assisted interpretation

Inborn errors of immunity (IEI) are rare monogenic diseases characterized by immune system dysfunctions. Molecular diagnostic methods, such as next-generation sequencing (NGS), are increasingly used for molecular diagnosis and exploring the genetic basis of these disorders. However, patients often receive inconclusive diagnoses due to variants of uncertain significance (VUS). Because of the high prevalence of VUS, reclassifying and prioritizing these variants for experimental validation is vital to optimize resources and improve diagnostic accuracy. Current variant classification tools follow recommendations from the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP), but they do not include structural descriptors. To address these issues, we created a dashboard to reclassify and prioritize VUS through a computational approach that combines biophysics, artificial intelligence, and functional data, enabling stratification of potentially pathogenic VUS for functional assays. We developed ProtSAVI (Protein Structure-Assisted Variant Interpretation), an R-based dashboard that offers a transparent, user-friendly framework for germline missense variant interpretation and classification. The tool integrates ACMG/AMP criteria, literature data, functional annotations, in silico fpredictions, and structural analyses to interpret and reclassify genetic variants. To evaluate our tool, we initially applied it to a dataset of 5 genes associated with autoinflammatory IEIs (ADA2, CDC42, STING1, TBK1, WDR1). Among the 662 variants listed as VUS in ClinVar over the past 10 years, we provided information for 82% of cases and reclassified 37.75% of this subset. Our dashboard aims to bridge the gap between clinicians and structural biologists and improve genetic diagnosis. It enhances traditional variant classification by incorporating structural insights into certain ACMG/AMP evidence-based criteria. The tool is scalable and applicable to analyze germline variants in other disease-related genes. Future versions will include additional descriptors, such as the effects of protein functional movements.