Molecular Outcome, Prediction, and Clinical Consequences of Splice Variants inCOL1A1, Which Encodes the proα1(I) Chains of Type I Procollagen

Approximately 10%-20% of germline pathogenic variants alter mRNA splicing, with phenotypes often dependent on the stability of the mRNA produced by the mutant allele. To better understand the relationships between genotype, mRNA splicing, and phenotype, we examined clinical and molecular data from 243 probands with osteogenesis imperfecta (OI) representing 145 unique splicing variants within the type I procollagen gene, COL1A1. All individuals with IVSX-1G>A mutations had OI type I because the substitution shifted the splice acceptor site 1 nt downstream and destabilized the mRNA. OI phenotypes were not consistent for any other splice variant identified. We sequenced all cDNA species from cultured dermal fibroblasts from 40 individuals to identify splice outcome and compared those results to splice predictions from Human Splice Finder (HSF), Spliceport (SP), and Automatic Splice Site and Exon Definition Analyses (ASSEDA). Software-based splice predictions were correct in 42%, 55%, and 74% instances for HSF, SP, and ASSEDA, respectively. As molecular diagnostics move increasingly to DNA sequence analysis, the need to understand the effects of splice site variants will increase. These data demonstrate that caution must be exercised when using splice prediction software to predict splice outcome.