Massively Parallel Characterization and Prediction of Disease-Associated Branch Site Variants that Perturb RNA Splicing

It is estimated that 10-30% of disease-associated genetic variants affect splicing. Splicing variants may generate deleteriously altered gene product and are potential therapeutic targets. However, experimental diagnosis for splicing variants is time-consuming and reliable computational prediction tools have not been established, especially for the 3' end of introns. The major challenge lies in the redundant and ill-defined branch site motif therein. Here, we carried out unbiased massively parallel splicing assays on 5,307 disease-associated variants overlapped with branch sites. We observed that 11.0% (455 out of 4,154 valid comparisons) of candidate variants showed a consistent pattern of altered splicing across four experimental replicates, among which 244 candidates (6.1%) presented more than two-fold changes in the use of noncanonical splice sites and these are named high-confidence (HC) significant candidates. Overall design: A total of 5,307 BS variants are reported in the Human Gene Mutation Database (HGMD), the public archive of interpretations of clinically relevant variants (ClinVar), the database of single nucleotide polymorphisms (dbSNP), or the Catalogue of Somatic Mutations in Cancer (COSMIC), all of which provide overlapping data on the most informative BS position 0 (i.e., the branchpoint), as well as the -2 or -3 positions of experimentally discovered branchpoints. We synthesized paired oligos containing the wildtype (WT) or mutant (mt) BS and their affiliated exons in bulk, ligated them into three-exon splicing minigenes, and then transfected them into human embryonic kidney cells (HEK293T) to examine the splicing outcome. An exonic barcode was associated with the intronic variant for genotype identification. The spliced product was recovered by reverse transcription and library-specific amplification, and then resolved by amplicon sequencing. Biased splicing efficiency and use of noncanonical splice sites by variants relative to their WT counterparts was identified by Fisher's exact test.