Variant Interpretation Pipeline Based on Large Language Models: Application to the PS3 Criterion for <i>MYH7/RYR1</i>

Genetic variant discovery has expanded dramatically with the advent of next-generation sequencing (NGS).Variant classification is both challenging and critical for clinical decision-making.The American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) established guidelines in 2015 to standardize variant classification [1].However, because the original guidelines are broad, disease-and gene-specific adaptations have been developed.These guidelines reflect the complexity of variant interpretation, as each criterion may vary across genes, and even a single criterion for a specific gene may be assigned different levels of strength depending on the available evidence [2].To address these challenges and support variant interpretation, various artificial intelligence (AI)-