Short tandem repeat variants are possibly associated with RNA secondary structure and gene expression

Short tandem repeats (STRs) are abundant in the human genome with approximately 300,000 embedded in gene introns, exons, and untranslated regions. High penetrance STR variants cause human diseases such as Myotonic dystrophy, Baratela-Scott syndrome, and various ataxias. The possibility that STRs contribute to polygenic disease is supported by recent high-powered datasets that link STRs to more subtle effects on gene expression. Indeed, STR variants can induce Z-DNA and H-DNA folding; alter nucleosome positioning; and change the spacing of DNA binding sites. On the other hand, little is known about how STR variants affect RNA secondary structure and accessibility. These factors could affect rates of splicing, nuclear export, and translation. We hypothesize that effects on RNA structure can be predicted using computational tools and associated with gene expression using DNA and RNA sequencing data. We test this hypothesis using data from the 1000 Genomes Project and ViennaRNA. We identify 17,255 transcribed STRs that affect RNA folding (fSTRs); 356 are possibly associated with gene expression. We characterize fSTRs by repeat motif, length, and gene level annotation. Transcribed fSTR variants tend to affect RNA multiloops and external loops. Effects on RNA accessibility depends on the repeat motif: a surprising result that is checked against simulation. These results shed light on how transcribed STRs affect RNA structure and pave the way for experimental validation.