P-09

Benjamin J. Capodanno1, Jeremy Stone1, Estelle Y. Da2, Sally B. Grindstaff1, David Reinhart1, Abbye E. McEwen1, Andrew B. Stergachis1, Lea M. Starita1, Douglas M. Fowler1, Alan F. Rubin2

1 University of Washington, Seattle, WA, USA
2 University of Melbourne, Melbourne, VIC, Australia

Variants of uncertain significance (VUS) complicate genomic medicine implementation because they have an unknown relationship to disease and cannot be used for clinical decision-making. While evidence from multiplexed assays of variant effect (MAVEs) can help resolve VUS, a major barrier to routine clinical use is that MAVE data have historically been locked in experiment-specific formats that downstream interpretation systems cannot consume without extensive manual intervention. MaveDB addresses this by implementing GA4GH data standards as the foundation of its data model and evidence-sharing infrastructure. Every variant is represented using VRS, enabling us to identify the same variant across independent experiments. For example, the TP53 variant NP_000537.3:p.Glu11Gln was submitted as part of six independent studies against three different experimental target sequences. Through VRS normalisation, we can easily resolve this to a single shared allele representation across all contributing datasets and connect these measurements to genomic variant data in other resources. VRS variants form the basis of structured functional measurements and clinical evidence, represented using VA-Spec statements across three layers of assertions: the numeric assay score, the variant consequence (e.g. functionally normal or abnormal), and the clinical evidence code (e.g. PS3_Strong or BS3_Moderate). Since clinical workflows typically operate on genomic nucleotide coordinates, for MAVE variants assayed only at the amino acid level, we use Cat-VRS categorical variant objects to transparently associate these to all possible underlying nucleotide variants while preserving the assay context. We have applied this implementation across 476,076 variant effect measurements from 82 MAVE datasets spanning 39 disease-associated genes, demonstrating how GA4GH standards can transform experimental functional data into computable evidence ready for clinical variant classification at scale.