Jessica White, MS
Assessing How Clinically Relevant Missense Mutations Impact Kinase Activity
Abstract
Research demonstrates that altered mutant kinase ATP affinity contributes to increased pathogenicity and inhibitor resistance, though systematic pan-kinome evaluation remains unexplored.(9-11) While characterization data for non-recurrent missense mutations on endogenous ligand affinity is limited, substantial KI affinity data exists for wild-type and recurrent mutant kinases. Since kinase sequences are relatively conserved, especially in the ATP-binding pocket, they are often viewed as mutant homologs of one another.(12) Relatedly, orthosteric, ATP-competitive compounds represent over 90% of currently approved KIs and are often designed to mimic endogenous ligands and exploit this conserved structure(13-16) Here, we describe a project combining machine learning (ML) techniques with experimental validation to assess the transfer learning potential of existing, large-scale kinase inhibitor datasets to predict the impact of missense mutations on endogenous kinase properties.
Kinase inhibitors are usually only tested in patient subgroups with recurrent mutations. This fellowship enables me to combine sequencing data and machine learning to assess the druggability of sporadic, collectively prevalent missense mutant kinases.
