Joint work with Wesley Tansey and the med-school-bound Amy Xie.

This work, as with all of our work, is the product of a community. It's the collective result of the bravery of our patients, of the many conversations (with our colleagues, trainees) and collaborations (with our clinicians), and the power and joy of revisiting scientific dogma

Those re-classified VUSs have significant clinical implications. Believe it or not: biallelical loss of (VUS or driver) KEAP1 is a predictive biomarker of inferior response in LUAD. Monoallelic loss (VUS or driver!) effectively WT. Zygosity itself matters.

Selection for biallelic inactivation is also translationally useful for classifying variants of unknown significance. Look at all those VUSs in KEAP1 that are functional, as validated by a base editing screen with Francisco Sanchez-Rivera and Sam Gould

Selection for biallelic inactivation is itself a highly useful metric for identifying cryptic driver mutations. APC is a rare driver in lung and prostate cancers

Quantifying selective pressure for biallelic inactivation, we found that genes assorted in four classes based on the prevalence of selection across diseases. Class I universally selected, but many of our favorite genes (ARID1A, PIK3R1) in Class 2/3/4

We provide a foundational but missing piece of data for the field at large: a resource of how often, and in what manner, genes are biallelically inactivated. Couldn't google it before, now you can. There's lots of heterogeneity across genes, and also within genes across diseases.

The goal was to assess the selective pressure for, and functional consequences of, biallelic activation in TSGs across our prospective clinical sequencing cohort. Unexpectedly (and so therefore, expectedly), things were not as expected. Two hit hypothesis be damned!