In panels A-C, all txrevise QTLs from promoters, internal exons and 3ʹ ends have been pooled to facilitate comparison with eQTLs as well as Leafcutter and full-length transcript usage QTLs. (A) Number of QTLs detected by read count, full-length transcript usage, Leafcutter and txrevise methods in each condition (N, naive; I, IFNɣ; S, Salmonella; I + S, IFNɣ+Salmonella; Ctrl; AcLDL) at 10% FDR. The number of QTLs detected by Leafcutter and txrevise are reported at the level of independent events (intron clusters or promoters/internal exons/3ʹ ends) and can include multiple QTLs per gene (Figure 2—figure supplement 2). The quantile-quantile plots are presented in Figure 2—figure supplement 3. (B) Sharing of QTLs detected by four quantification methods. The numbers on the heatmap show the fraction of QTLs detect by one method that were replicated by each of the three other methods (r2 >0.8 between lead variants). Only QTLs with FDR < 0.01 were included in the analysis. (C) Enrichment of genomic annotations at QTLs detected by the four quantification methods. (D) Comparison of Leafcutter tuQTLs to promoter, internal exon and 3ʹ end usage QTLs detected by txevise. Genomic annotations used for enrichment analysis: promoter - promoter flanking regions (−2000 bp to +200 bp); 5ʹ UTR, coding, intron, 3ʹ UTR - corresponding regions extracted from Ensembl transcripts; poly(A) - experimentally determined polyadenylation sites (±25 bp) (Gruber et al., 2016); open chromatin - open chromatin regions from macrophages (Alasoo et al., 2018); splicing factor - experimentally determined binding sites of splicing factors detected by eCLIP (Van Nostrand et al., 2017). The points on panels C and D show the natural logarithm of enrichment for each annotation and the lines represent the 95% confidence intervals from fgwas (Pickrell, 2014).