AAVC: An automated framework for high-accuracy ACMG-based variant classification

Abstract

Classification of DNA sequence data requires the implementation of the American College of Medical Genetics and Genomics (ACMG) standards and guidelines. Therefore, automated tools have been developed. However, these tools often lack robust and up-to-date methodologies. This study reports on the development of a new tool and examines its performance for diagnostic and research purposes. Methods The automated ACMG-based variant classifier (AAVC) presented here computationally analyzes sequence variants following the ACMG guidelines, the Clinical Genome Resource specifications and a novel framework by leveraging large public databases and in silico prediction tools. Results AAVC demonstrated high concordance (94.39%) with the Food and Drug Administration recognized variant classifications, outperforming currently available tools. It classified 55% of the variants of uncertain significance in clinical variation into clinically significant categories. We identified, in the Turkish Variome, 215 novel pathogenic, likely pathogenic, or variants of uncertain significance high variants in the secondary finding genes and revealed that 1 in 10 individuals carried an actionable genotype. Conclusion AAVC constitutes a robust framework for the accurate classification of human germline sequence diversity is available at https://aavc.bilkent.edu.tr/, offering a highly accurate, rapid, and up-to-date platform for clinical laboratories and research groups to automatically interpret sequence variants.