A multiplexed assay for exon recognition reveals that an unappreciated fraction of rare genetic variants cause large-effect disruptions to splicing

Mutations that cause exon skipping can have severe consequences on gene function and cause disease. Here we explore how human genetic variation affects exon recognition by developing a Multiplexed Functional Assay of Splicing using Sort-seq (MFASS). We assayed 27,733 variants in the Exome Aggregation Consortium (ExAC) within or adjacent to 2,198 human exons in the MFASS minigene reporter, and found that 3.8% (1,050) of variants, most of which are extremely rare, led to large-effect splice-disrupting variants (SDVs). Importantly, we find that 83% of SDVs are located outside of canonical splice sites, are distributed evenly across distinct exonic and intronic regions, and are difficult to predict a priori. Our results indicate extant, rare genetic variants, even outside the context of disease, can have large functional effects at appreciable rates, and that MFASS enables their empirical assessment for large-effect splicing defects at scale. MFASS is a splicing minigene reporter. There are three synthetic oligo libraries, SRE library in two backbones and the SNV library. The oligo library is cloned in between a split GFP reporter with a downstream mCherry reporter. Exon inclusion leads to both GFP and mCherry fluorescence, exon skipping leads to only mCherry fluorescence. We sort the libray of cells based on these two fluorescent reporters using fluorescence activated cell sorting (FACS). We then sequence each individual bin to get a quantitiative measure of splicing. For the SRE library, we sort into three bins, with two replicates. Double positive (DP) indicates both GFP and mCherry are present and correspond to exon skipping. Intermediate (INT) indicates intermediate levels of GFP and mCherry. Single positive (SP) indicates no GFP present but mCherry present, corresponds to exon inclusion. For the SNV library, the sorting is similar except there are four bins, also done in replicate. Hi indicates both GFP and mCherry, Int-Hi and Int-Lo correspond to two intermediate bins, and Lo corresponds to no GFP present but mCherry present. We use the read counts across all bins, present in processed data files, to calculate an inclusion index for an individual sequence.