Beyond estimating overall levels of gene expression, RNA-seq data also allow for the quantification of expression levels of individual transcript isoforms, as well as components of these, such as exons, splice junctions, and untranslated regions. We refer to the quantitative variation of gene structure due to genetic variation as splicing QTLs (sQTLs) (Fig. 5A). To identify sQTLs, we used Altrans (35), a method that identifies SNPs that are associated with variation in the expression levels of exon junctions (sjQTLs; box S1) (14), and sQTLSeekeR (36), a method that identifies SNPs associated with the variation in the relative abundances of gene transcript isoforms (srQTLs; box S1) (14). Altrans identifies both novel and annotated splicing events, while sQTLSeekeR tests only annotated isoforms. Altrans, however, is restricted to changes in the usage of splice junctions, while sQTLSeekeR can in principle detect any variation in the relative abundance of transcript isoforms (fig. S23). Altrans was run using a ±1 Mb region around the TSS, while for sQTLseekeR, we tested within the body of the gene ±5 kb (14).
