Phenome-wide identification of therapeutic genetic targets, leveraging knowledge graphs, graph neural networks, and UK Biobank data.

To test whether Mantis-ML can improve triaging among the highly ranked but not statistically significant genetic associations or among the many genes that could be driving a common variant loci signal, we cross-referenced genes with Mantis-ML predictions to associate with a trait against genes nominally significant ( P < 0.05) for the trait in a recent phenome-wide association study (PheWAS) of 454,669 exomes from the UK Biobank (UKB) ( 2 ).

also in 7,732 other papers