Evaluating the clinical application of computational methods for identifying splice variants in cardiomyopathy-associated genes
OBJECTIVE:Non-canonical splice variants are increasingly recognized as key contributors to inherited cardiomyopathies. Several computational methods have been developed to prioritize the effects of variants on splicing. However, their application in the genetic diagnosis of cardiomyopathies remains unclear. Thus, the performance of computational methods for cardiomyopathy-associated genes and of their clinical applications in hypertrophic cardiomyopathy (HCM) patients were evaluated. METHODS:Whole-exome sequencing was performed on 1212 unrelated patients with HCM, and splice-site variants in cardiomyopathy-associated genes were collected from scientific publications. A splicing-region-specific strategy was employed to evaluate performance of the computational methods and to used optimal methods for prioritizing candidate splice variants. Splicing effects of these variants were assessed using in vitro Minigene assays. RESULTS:The performance evaluation showed that the optimal methods were Pangolin (deep exonic,core donor,and extended donor regions), MLCsplice (extended donor, core acceptor, and extended acceptor regions), MMSplice, and SpliceAI (extended donor region). Using these optimal methods, 53.5％ (23/43) of the variants were prioritized and confirmed to cause aberrant splicing. A total of 23 splice variants were found in 4.5％ (54/1212) of the unrelated HCM cases. CONCLUSION:Based on the splicing-region- specificity, the prediction strategy effectively identified splice variants, thereby improving the diagnostic value of genetic testing for inherited cardiomyopathies.
万方数据
维普
