Extended Data Fig. 4: Per-junction and per-gene Manhattan plots for splicing efficiency and gene expression QTL.

Manhattan plots showing splicing efficiency and whole-transcript expression QTL effects (P-values calculated using 2-sided t-tests). Splicing efficiency analysis was performed for the FGD4 (a) and GALNT2 (b) genes, with associations highlighting junctions for which splicing appears to be genetically modulated in cis. The proportion of spliced to un-spliced reads at each junction has been treated as an individual phenotype, with association analysis performed using intervals of imputed sequence data spanning the annotated gene structures, and ± 100kbp 5’ and 3’ of the gene boundaries. The splicing junction predicted to be impacted by the splice donor (FGD4) and acceptor (GALNT2) mutations is indicated by the blue highlighted panels, with the candidate causative mutation indicated in red. Whole transcript eQTL analysis was performed to assess possible gene expression impacts as a consequence of non-sense mediated RNA decay (NMD) for the FGD4 (c), GALNT2 (d), and MUS81 (e) genes that harbour non-sense mutations. Note that for the two genes that show significant eQTL (FGD4 and MUS81), the mutant allele is overexpressed and thus no NMD is apparent. In the case of the MUS81 and GALNT2 genes, lack of apparent NMD can be anticipated given the position of the GALNT2 c.1561-1G>A mutation in the final exon, and the presence of an in-frame start codon (p.Met76) following the p.Gly70 mutation for MUS81.