Splice Site Variants in the and Genes: Transcript Analysis as a Tool in Supporting Pathogenicity.

Background: Approximately 75% of clinically definite long QT syndrome (LQTS) cases are caused by mutations in the , and genes. Of these mutations, a small proportion (3.2-9.2%) are predicted to affect splicing. These mutations present a particular challenge in ascribing pathogenicity.

Methods: Here we report an analysis of the transcriptional consequences of two mutations, one in the gene (c.781_782delinsTC) and one in the gene (c.2437-5C>A), which are predicted to affect splicing. We isolated RNA from lymphocytes and used a directed PCR amplification strategy of cDNA to show mis-spliced transcripts in mutation-positive patients.

Results: The loss of an exon in each mis-spliced transcript had no deduced effect on the translational reading frame. The clinical phenotype corresponded closely with genotypic status in family members carrying the splice variant, but not in family members with the splice variant. These results are put in the context of a literature review, where only 20% of all splice variants reported in the , and gene entries in the HGMDPro 2015.4 database have been evaluated using transcriptional assays.

Conclusions: Prediction programmes play a strong role in most diagnostic laboratories in classifying variants located at splice sites; however, transcriptional analysis should be considered critical to confirm mis-splicing. Critically, this study shows that genuine mis- splicing may not always imply clinical significance, and genotype/phenotype cosegregation remains important even when mis-splicing is confirmed.