Identifying deleterious noncoding variation through gain and loss of CTCF binding activity.

Abstract

CCCTC binding factor (CTCF) regulates gene expression through DNA binding at thousands of genomic loci. Genetic variation in these CTCF binding sites (CBSs) is an important driver of phenotypic variation, yet extracting those that are likely to have functional consequences in whole-genome sequencing remains challenging. To address this, we develop a hypothesis-driven framework to identify and prioritize CBS variants in gnomAD. We synthesize CTCF's binding patterns at 1,063,878 genomic loci across 214 biological contexts into a summary of binding activity. We find that high binding activity significantly correlates with both conserved nucleotides (Pearson R = 0.35, p < 2.2 × 10^-16) and sequences that contain high-quality CTCF binding motifs (Pearson R = 0.63, p = 2.9 × 10^-12). We then use binding activity to evaluate high-confidence allelic binding predictions for 1,253,329 single-nucleotide variations (SNVs) in gnomAD that disrupt a CBS. We find a strong, positive relationship between the mutability-adjusted proportion of singletons (MAPS) metric and the loss of CTCF binding at loci with high in vitro activity (Pearson R = 0.74, p < 2.2 × 10^-16). To contextualize these findings, we apply MAPS to other functional classes of variation and find that a subset of 339,380 loss of CTCF binding variants is observed as infrequently as missense variants are. This work nominates these thousands of rare, noncoding variants that disrupt CTCF binding for further functional studies while providing a blueprint for prioritizing variation in other transcription factor binding sequences.