A Framework for Variant Filtering to Improve Diagnostic Outcomes in Genome Sequencing

Rare genetic disorders frequently remain undiagnosed despite the use of exome sequencing (ES) and genome sequencing (GS). This study aimed to enhance diagnostic yield by introducing a systematic framework: SMART filtering (Systematic, Multi-step, Automated, Rare variant prioritization Tool), designed to optimize GS analysis, with a particular emphasis on non-coding regions. We retrospectively analyzed 153 cases with previously negative ES and GS results. SMART filtering was applied as a pre-annotation filtering step, integrating conservation metrics, allele frequency thresholds, and splicing prediction scores. Then followed by phenotype-based prioritization and manual review. Application of this framework identified pathogenic or likely pathogenic variants in 4.6% of cases, strong variants of uncertain significance (VUS) in 1.3%, and additional candidate variants in 5.2% of cases. Several deep-intronic variants were confirmed by RNA studies and segregation analyses. The framework enabled efficient prioritization of non-coding and regulatory variants that were overlooked by standard analyses. In conclusion, SMART filtering offers a reproducible and scalable strategy to improve diagnostic yield in GS, particularly in cases where conventional analysis methods fall short. Integrating such frameworks into clinical genomic pipelines can enhance variant interpretability and diagnostic performance, ultimately leading to better patient care outcomes.