A systematic assessment of the impact of rare canonical splice site variants on splicing using functional and in silico methods

Background/ObjectivesCanonical splice site variants (CSSVs) are often presumed to cause loss-of-function (LoF) and are assigned very strong evidence of pathogenicity (according to ACMG criterion PVS1). However, the exact nature and predictability of splicing effects of unselected rare CSSVs in blood-expressed genes is poorly understood. MethodsA total of 184 rare CSSVs in unselected blood-expressed genes were identified by genome sequencing in 121 individuals, and their impact on splicing was interrogated manually in RNA sequencing (RNA-seq) data. Blind to these RNA-seq data, we attempted to predict the precise impact of CSSVs by applying in silico tools and the ClinGen Sequence Variant Interpretation Working Group 2018 guidelines for applying PVS1 criterion. ResultsThere was no evidence of a frameshift nor of reduced expression consistent with nonsense-mediated decay (NMD) for 24% of CSSVs: 17% had wildtype splicing only and normal junction depths, 3.25% resulted in cryptic splice site usage and in-frame indels, 3.25% resulted in full exon skipping (in-frame), and 0.5% resulted in full intron inclusion (in-frame). Misclassification rates for splicing outcome (frameshift/NMD vs. no frameshift/no NMD) using (i) SpliceAI, (ii) MaxEntScan, and (iii) AutoPVS1 ranged from 30-41%, with none outperforming a simple "zero rule" classifier. ConclusionNearly 1 in 4 CSSVs may not cause LoF based on analysis of RNA-seq data. Predictions from in silico methods were often discordant with findings from RNA-seq. More caution may be warranted in applying PVS1-level evidence to CSSVs in the absence of functional data.