A method to predict the impact of regulatory variants from DNA sequence

Abstract

Most variants implicated in common human disease by Genome-Wide Association Studies (GWAS) lie in non-coding sequence intervals. Despite the suggestion that regulatory element disruption represents a common theme, identifying causal risk variants within indicted genomic regions remains a significant challenge. Here we present a novel sequence-based computational method to predict the effect of regulatory variation, using a classifier (gkm-SVM) which encodes cell-specific regulatory sequence vocabularies. The induced change in the gkm-SVM score, deltaSVM, quantifies the effect of variants. We show that deltaSVM accurately predicts the impact of SNPs on DNase I sensitivity in their native genomic context, and accurately predicts the results of dense mutagenesis of several enhancers in reporter assays. Previously validated GWAS SNPs yield large deltaSVM scores, and we predict novel risk SNPs for several autoimmune diseases. Thus, deltaSVM provides a powerful computational approach for systematically identifying functional regulatory variants.

Insight into GATA1 transcriptional activity through interrogation ofciselements disrupted in human erythroid disorders

“…To address this, we coupled experimental and bioinformatics approaches to show that mutations in GATA1 and CF BSs can substantially affect genes implicated in MEDs, and created mutation maps of predicted mutations across CREs proximal to MED genes (15,16). These maps may prove useful for prioritizing variants from WGS or targeted sequencing of MED cases, as we have initially shown for PKLR-RE1 mutations that disrupt a TAL1 binding motif, and can identify disruptive as well as gain-of-function mutations.…”

“…ChIP-seq and RNA-seq were processed as described previously (32). Random forests were used to model expression, k-means and PAM were used for clustering, and gkmer-SVM and DeepBind were used to create mutation maps (15,16). Complete details are available in SI Materials and Methods.…”

“…This difficulty is most clearly reflected in the distribution of mutations listed in databases of Mendelian disorders, such as the Human Gene Mutation Database, where most mutations are found within coding regions (86%) or at intronic splice sites (11%), with only a small fraction (3%) identified in regulatory regions (14). Newer methods for annotating and predicting the impact of noncoding (NC) variants have provided substantial improvements (15,16), but experimental validation of the presumed effects remains critical for the determination of pathogenicity and elucidation of the mechanism of action (13,17).…”

“…First, we trained a gapped k-mer support vector machine (gkmer-SVM) on EB and K562 open chromatin data and used delta-SVM to predict single nucleotide effects (16). We then used already-trained models for the TFs GATA1, TAL1, KLF1, and NFE2 from DeepBind, a convolutional neural network approach (15).…”

A genome-wide interactome of DNA-associated proteins in the human liver

“…We computed a "delta" binding score for all possible point mutations within DAP binding sites, defined as the mean decrease in our SVMs' classifier value for the alternate base relative to the reference sequence. This strategy is similar to a previously developed approach, "deltaSVM," that focuses on local disruptions of 10-mer feature weights (Lee et al 2015). Bases with the most negative delta binding score tended to be the most highly conserved for most DAPs (Supplemental Table S10).…”