Genome editing of noncoding regions to map sequence-function relationships at scale

Deciphering how genetic variants lead to disease can advance precision diagnostics and mechanistically informed therapies. Yet, our ability to link variants in the human genome’s vast noncoding regions to phenotypic consequences is highly limited. This stems largely from a lack of experimental evidence on individual variants’ functional effects. Here, we propose to develop and deploy a suite of new genome editing methods to generate functional data for noncoding variants at unprecedented scale.First, we will leverage a prime editing screening platform we have recently established to systematically ask how intronic variants and 5’-untranslated region (5’UTR) variants lead to loss of function (LoF) across dozens of human disease genes. Second, we will interrogate relationships between regulatory sequence variants, transcription factor binding sites (TFBSs), and genomic contexts across diverse cell types using novel methods to link variants to effects on messenger RNA (mRNA) levels. Third, we will uncover new noncoding variants underlying disease by combining variant effect prediction with large-scale functional assessment.Collectively, this work will establish new technologies to experimentally characterize noncoding variants, delineate sequence-function relationships from 100,000s of variant effect measurements, and substantially advance our ability to predict, identify, and engineer noncoding variants of functional importance.