Many long intergenic non-coding RNAs distally regulate mRNA gene expression levels

Abstract

Long intergenic non-coding RNAs (lincRNA) are members of a class of non-protein-coding RNA transcript that has recently been shown to contribute to gene regulatory processes and disease etiology. It has been hypothesized that lincRNAs influence disease risk through the regulation of mRNA transcription [88], possibly by interacting with regulatory proteins such as chromatin-modifying complexes [37, 50]. The hypothesis of the regulation of mRNA by lincRNAs is based on a small number of specific lincRNAs analyses; the cellular roles of lincRNAs regulation have not been catalogued genome-wide. Relative to mRNAs, lincRNAs tend to be expressed at lower levels and in more tissue-specific patterns, making genome-wide studies of their regulatory capabilities difficult [15]. Here we develop a method for Mendelian randomization leveraging expression quantitative trait loci (eQTLs) that regulate the expression levels of lincRNAs (linc-eQTLs) to perform such a study across four primary tissues. We find that linc-eQTLs are largely similar to protein-coding eQTLs (pc-eQTLs) in cis-regulatory element enrichment, which supports the hypothesis that lincRNAs are regulated by the same transcriptional machinery as protein-coding RNAs [15, 80] and validates our linc-eQTLs. We catalog 74 lincRNAs with linc-eQTLs that are in linkage disequilibrium with TASs and are in protein-coding gene deserts; the putative lincRNA-regulated traits are highly enriched for adipose-related traits relative to mRNA-regulated traits.