A framework for interpreting global effects of genetic variants contributing to disease risk - advances in next-generation sequencing are enhancing the routine detection of human genetic variants,…

A framework for interpreting global effects of genetic variants contributing to disease risk - advances in next-generation sequencing are enhancing the routine detection of human genetic variants, particularly in clinical settings. Yet, the ability to interpret the functional consequences of these variants has lagged far behind. While the identification of clinically actionable and pathogenic mutations has revolutionized the field of precision medicine, these unfortunately represent a small minority of reported human genetic variants. A large fraction of patients (~30-70%) who undergo diagnostic genome sequencing are found to have variants of unknown significance (VUS), for which a clinical impact cannot be assigned. Multiple computational methods have been designed to score the severity of a particular protein-coding mutation. While informative, predictions from these methods are imperfect and do not give mechanistic insights into a variant's impact. On the other hand, laboratory-based functional genomics approaches have been used to characterize individual pathogenic variants in human cell lines or model organisms, but these methods are low-throughput and can only focus on a handful of mutations. Given the sheer volume of identified (and as-yet undiscovered) genetic variants in need of clinical interpretation, there is a pressing need for high-throughput technologies to address this challenge. Here, we describe complementary experimental and computational strategies for high-throughput characterization of the impact of genetic variants in key regulatory and DNA repair genes. We focus on mutations in "trans-acting factors" such as chromatin regulators, transcription factors and DNA repair genes, which are widely implicated in human disease. Mutations in these genes have the potential to induce widespread transcriptomic, chromatin or genomic changes, and thus are most amenable to the strategies described in our proposal. Notably, while these categories represent only a subset of clinically actionable genes, they nonetheless encompass thousands of potential gene targets. We interrogate two classes of phenotypes: global transcriptomic/chromatin changes induced by mutations to key transcription factors or other regulatory proteins, and mutator phenotypes induced by disrupting proteins involved in DNA repair processes. We first develop scbe-SEQ (single cell base editor sequencing), which combines pooled, high-precision genome editing with single-cell sequencing assays to interrogate effects of hundreds of variants simultaneously. Scbe-SEQ leverages chemical base editing (be) to introduce specific single nucleotide variants (snvs). In parallel, we develop computational approaches leveraging genomics datasets from large biobanks to identify genes harboring high- impact variants which can be followed up using scbe-SEQ. Overall, our proposal brings together complementary expertise spanning computational human genomics, molecular biology, genome editing and development/implementation of genomic technologies. We envision the proposed experimental and computational strategies will make important advances in the ability to interpret the impact of individual variants in a variety of contexts and provide insights for predicting the effects of previously unobserved genetic variants.