Fig. 4
From: Clinical implementation of RNA sequencing for Mendelian disease diagnostics
Aberrant splicing. A Distribution of genes per sample that had at least one splicing outlier, for all genes and genes known to cause disease (OMIM). B Observed over expected number of splicing outliers on different gene categories. Neuroblastoma breakpoint family (NBPF) and collagen genes were chosen due to their high number of exons and due to collagen genes alternative splicing in a developmental-stage manner and NBPF genes having a repetitive structure, which exposes them to illegitimate recombination. Error bars represent 95% confidence intervals of pairwise logistic regressions. C Split-read counts (y-axis, gray junction on panel E) of the first annotated junction of TWNK against the total split-read coverage (x-axis, gray and red junctions on panel E) of the first donor site of TWNK. Many samples are not exclusively using the annotated junction (scattered below the diagonal), leading to a reference 𝜓5 for the annotated junction of 87%. The observed 𝜓5 for the first acceptor site of TWNK in the outlier sample is 20% (obtained by dividing the junction reads by the total junction coverage, 4/20). D Gene-level significance (−log10(P), y-axis) versus differential splicing effect (observed minus expected usage proportion of the tested donor site, Δ𝜓5, x-axis) for the alternative splice donor usage in sample R36605. Gene-level significance was obtained after multiple-testing correction across junctions. Outliers are marked in red and the gene TWNK is explicitly labeled. The Δ𝜓5 value for the first donor site of TWNK in this sample is − 0.67 = 0.2–0.87. E Schematic depiction of the synonymous NM_001163812.1:c.1302C>G (p.Ser434=) TWNK variant and its consequence on the RNA level, activating a new acceptor site (ACGG in red) and leading to the creation of a premature termination codon (red rectangle) in sample R36605. The percentage of each transcript isoform is shown next to it. Figure not shown at genomic scale