Abstract:Genome-wide association studies (GWAS) have linked thousands of genetic variants to various complex traits or diseases. However, most identified variants have weak individual effects, are correlated with nearby polymorphisms due to linkage disequilibrium (LD), and are located in non-coding cis-regulatory elements (CREs). These characteristics complicate the assessment of the direct impact of each variant on tissue specific gene expression and phenotype. To address this challenge, we have developed a novel algorithm that leverages polymer folding and 3D chromatin interactions to prioritize and identify putative causal variants and their target genes. From the millions of eQTL-Gene pairs identified by GTEx in human somatic tissues, we classify only ~10-20% as putative functional eQTL-Gene pairs supported by phenotypic associations confirmed through CRISPR deletion experiments. Our findings show that unlike most variants, functional eQTL-Gene pairs predominantly reside within the same topologically associating domain (TAD) and have strong associations with cell-type specific cis-regulatory elements (CREs), enriched for binding sites of tissue-specific transcription factors. Unlike most approaches that rely on linear distance or other chromatin features (histone code, accessibility), our algorithm emphasizes the importance of physical interactions and 3D chromatin folding in gene regulation, as the identified eQTL-Gene pairs are all among the small fraction of physical chromatin interactions sufficient for chromatin locus folding. Overall, our algorithm reduces false positive associations between DNA variants and genes identified by eQTL analysis and uncovers novel variant-gene pair associations. These findings suggest a mechanism where a small number of regulatory variants control tissue specific gene expression via their physical association with target genes confined within the same TAD. Our approach provides new insights into the molecular mechanisms driving GWAS phenotypes.

Regulatory features aid interpretation of 3′UTR variants

Predicting functional UTR variants by integrating region-specific features

Quantifying negative selection in human 3ʹ UTRs uncovers constrained targets of RNA-binding proteins

Massively parallel screen uncovers many rare 3′ UTR variants regulating mRNA abundance of cancer driver genes

Human 5′ UTR design and variant effect prediction from a massively parallel translation assay

Deciphering 3'UTR Mediated Gene Regulation Using Interpretable Deep Representation Learning

Using Population-Scale Transcriptomic and Genomic Data to Map 3′ UTR Alternative Polyadenylation Quantitative Trait Loci

Regulatory variants: from detection to predicting impact

Genetic Basis of Alternative Polyadenylation is an Emerging Molecular Phenotype for Human Traits and Diseases

A New Approach for Discovering Functional Links Connecting Non-Coding Regulatory Variants to Gene Targets

Identification of candidate regulatory sequences in mammalian 3' UTRs by statistical analysis of oligonucleotide distributions

Dynamic Variations of 3′UTR Length Reprogram the mRNA Regulatory Landscape

Differences in 5'untranslated regions highlight the importance of translational regulation of dosage sensitive genes

Multiplexed Assays of Human Disease-relevant Mutations Reveal UTR Dinucleotide Composition as a Major Determinant of RNA Stability

Immune-response 3′UTR Alternative Polyadenylation Quantitative Trait Loci Contribute to Variation in Human Complex Traits and Diseases

A comprehensive analysis of 3′UTRs in Caenorhabditis elegans

Disrupting upstream translation in mRNAs is associated with human disease

A Comprehensive Analysis of 3’UTRs in

Massively Parallel Polyribosome Profiling Reveals Translation Defects of Human Disease-Relevant UTR Mutations

Association Between a Functional Variant at PTGS2 Gene 3′UTR and Its Mrna Expression in Lymphoblastoid Cell Lines

Deciphering cis-regulatory elements using REgulamentary