Approximately half of individuals with a suspected monogenic disorder remain undiagnosed after exome or genome sequencing. Functional RNA evidence can resolve variants that alter transcript splicing and abundance, but 1,436 Mendelian disease genes (~1/3 of all) are expressed too weakly in clinically accessible patient samples of blood or skin fibroblasts for standard RNA-seq analysis. Consequently, the >85,000 variants of uncertain significance in ClinVar predicted to affect RNA splicing or decay of these so-called lowly expressed disease genes (LDGs) lack a practical route to functional assessment.
We developed a targeted RNA Capture-Seq panel covering protein-coding isoforms of the 1,436 LDGs to enrich LDG RNA prior to RNA-Seq. We applied the assay to RNA from blood, blood-derived lymphoblastoid cell lines (LCLs), and human dermal fibroblasts, followed by 150-bp paired-end short-read sequencing to a depth of 100 million reads. The performance of RNA Capture for the analysis of LDG splicing was evaluated using splice-junction read depth and coverage of clinically relevant transcripts.
Custom LDG RNA Capture-Seq increased splice-junction read depth across targeted genes by >90-fold relative to non-capture RNA-seq or standard exome-wide RNA Capture-Seq. At a conservative threshold of at least 50 reads across more than 90% of splice junctions, 680 of 1,436 genes (47.3%) were deemed compliant for RNA-seq-based assessment of splicing in at least one accessible sample type. Furthermore, captured fibroblast transcripts from neurological disease genes showed concordant splicing with neuronal samples (>98% of 13,166 splicing events conserved), supporting the use of RNA from accessible tissues as a surrogate for clinically relevant tissues. We then applied RNA Capture to reveal the functional impact of variants of uncertain significance identified in LDGs HECW2, MED12L, JPH2, NHS, ACAN, and TBX1, supporting reclassification of variant pathogenicity. In one example, although SpliceAI predicted null effect of a MED12L c.1258G>T variant, LDG RNA Capture revealed it caused an out-of-frame exon-skipping, which was validated using orthogonal MED12L transactivation, and confirmed a loss-of-function effect.
Targeted LDG RNA Capture-Seq makes hundreds of previously inaccessible disease-gene transcripts amenable to functional analysis from routinely obtainable samples in a single, high-throughput and standardised method. This scalable, variant-agnostic assay could expand RNA-based interpretation of unresolved rare-disease variants while retaining the patient's genetic and cellular context.