Background: Finding a mutation in an ion channel gene in Long QT syndrome allows the accurate identification of other family members at risk of sudden death. Interpretation of genetic investigations is complicated due to the majority of mutations in Long QT syndrome being missense mutations, frequently novel, and difficult to distinguish from rare polymorphisms.
Method: We present a new model that combines an analysis of 5 independent areas: the proband’s clinical phenotype, familial segregation, evolutionary conservation, amino acid properties and data from control samples to generate an overall likelihood ratio that an observed variation is a pathogenic mutation. The ability of the model to improve the interpretation of DNA sequencing results was examined in 50 families with a clinical diagnosis of Long QT syndrome. Complete sequencing of six LQT genes (LQT1,2,3,5,6 and 7) was performed and the integrated model used to analyse any missense variants.
Results: Likelihood ratios strongly favouring pathogenicity were demonstrated for previously described mutations as well as novel mutations clinically suspected to be pathogenic. For recurrent mutations (e.g. KCNQ1 G314S) the LR exceeded 1000:1 in favour of causality while known polymorphisms analysed by the model showed LRs strongly against causation.
Conclusion: The described model provides an objective tool to analyse missense variants in long QT syndrome, allowing accurate interpretation of novel variants in a clinical time-frame and facilitating further of predictive testing in at risk families.