SeverYX: An ensemble framework for robust variant severity classification
Accurate interpretation of genetic variants in clinical genomics depends on reconciling multiple computational predictors that often disagree or provide incomplete evidence. We present SeverYX, a transparent, rule-based ensemble framework that consolidates conservation, splicing, and functional-impact predictors into a three-tier severity classification ('High', 'Medium', 'Low') through an…
An ensemble framework called SeverYX has been introduced for reliably classifying the severity of genetic variants in clinical genomics. The framework integrates various computational predictors that may sometimes provide conflicting or partial information, leading to the need for a unified classification system. SeverYX categorizes variants into three tiers of severity: High, Medium, and Low.
To verify the effectiveness of SeverYX, it was evaluated using a dataset of 270,208 variants from ClinVar. The dataset included 86,161 variants classified as pathogenic or likely pathogenic, as well as 184,047 variants classified as benign or likely benign, stratified by their molecular consequences, review status, and haploinsufficiency.
The baseline configuration of SeverYX achieved 97.6% concordance with the ClinVar dataset. At a fixed threshold for the High severity classification, it correctly assigned 90.0% of pathogenic variants and 2.0% of benign variants as High severity, indicating a strong agreement with the curated dataset. The Concordance was highest for loss-of-function variants and lowest for non-coding and unclassified pathogenic variants, reflecting the limitations of existing predictive tools in these areas.
By optimizing the thresholds and incorporating additional predictors such as AlphaMissense and REVEL as missense predictors, overall concordance improved to 98.7%. However, this adjustment came at the cost of reduced sensitivity for pathogenic variants, dropping to 88.3% for High severity classification. The Matthews correlation coefficient, a measure of the model's performance, increased from 0.895 to 0.908, demonstrating the effectiveness of SeverYX in consolidating heterogeneous evidence into clinically interpretable severity classifications.
The introduction of SeverYX represents a transparent and reproducible approach for integrating heterogeneous variant-effect evidence into clinical decision-making. Its extensible design allows for the seamless incorporation of future predictive models, further enhancing its utility in genomic interpretation workflows.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
