PhD Seminar: Daniel Tabet
PhD Oral Examination for Daniel Tabet
Date: Thursday, September 5th, 2024 at 10:00 AM
Location: Red Room, Donnelly Centre or via Zoom
Abstract:
Approaches to identifying pathogenic human missense variation—with application to the familial hypercholesterolemia gene LDLR
Achieving the promise of personalized medicine requires that we be able to diagnose disease and infer patient risk based on genetics. Unfortunately, the impact of genotype on a given phenotype cannot yet be readily (nor reliably) inferred from sequence. While population-based studies have succeeded in associating many common alleles with human phenotypes, individually these do not typically impact risk enough to alter clinical care. Rare alleles—which account for most genetic variation and tend to have larger effect sizes—have by comparison been difficult to associate with disease outcomes. Given both the pace and promise of genetic sequencing, there is a pressing need for scalable evidence sources that can inform rare variant interpretation. Here, two approaches have emerged: in vitro sequence–function studies which connect genotype to phenotype in an experimental model; and computational variant effect predictors which infer variant effect based on sequence properties. In this work, I present a comprehensive sequence–function study of coding variation in the familial hypercholesterolemia gene LDLR, characterizing sequence-structure function relationships and providing scaled evidence towards clinical variant classification and disease risk prediction. I also present a novel and unbiased method for benchmarking computational variant effect predictors using human phenotypes in population-based cohorts, which can serve as a readily adoptable method by which to inform predictor choice for personal and clinical genetics. Together, this work provides evidence and methodology that can be applied towards connecting variant to function at scale, and towards improving the accuracy and sensitivity of genetic medicine.