The performance of DeltaSplice and baseline methods in predicting splicing-altering de novo mutations associated with neurodevelopmental disorders (NDDs). (A) Enrichment of autism-associated DNMs in the WGS data set. Stop-gain/loss or splice-site mutations were excluded from this analysis. DNMs in autism patients or controls were combined and ranked by ΔPSIalt-ref. Among the top k predictions using varying k, the enrichment of autism-associated mutations is shown. (B) Similar to A, but for DNMs identified in NDD and control samples in the WES data set. (C) Genes with two or more splicing-disrupting mutations (as defined by ΔPSIalt-ref < −0.05). The number of likely gene-disrupting (LGD) mutations (stop-gain or splice-site mutations, but not frame-shifting indels) and SFARI autism gene scores is shown. NDD-risk genes predicted based on de novo mutation burdens, with or without including predicted splicing-altering mutations, are also indicated.
