A Mammalian Genomic Signature Shaped by Single Nucleotide Variants Controlling Transcriptome Integrity and Diversity

Many functional features of mammalian genomic sequences remain poorly understood. Here, we identify a widely evolved genomic signature, G-tract-AG motifs consisting of guanine tracts closely upstream of AG dinucleotides, which is significantly associated with single-nucleotide variants (SNVs) identified in genome-wide association studies (GWAS), particularly within non-coding regions (NCRs). Approximately 9,000 such G-tracts within human genes are disrupted by variants of the cis -splicing quantitative trait loci (sQTLs) in the Genotype-Tissue Expression (GTEx) project. Functionally, G-tracts repress splicing at the adjacent 3′AG, primarily by stalling the second transesterification step. Disruption of the G-tracts by SNVs relieves this repression, enabling splicing and generating novel transcript isoforms. These G-tract-disrupting SNVs are in cis across the majority of protein-coding genes and are among thousands of rare variants causing genetic diseases. Our findings provide mechanistic insights into the maintenance of transcriptome integrity while still allowing diversity by a mammalian-evolved genomic signature, particularly for the NCR SNVs in association with diverse traits and a new framework for their functional annotation.