Action and the Actionability in Exome Variation

As next-generation sequencing emerges as a clinical tool, one of the greatest challenges facing human geneticists and clinicians is the interpretation of the vast numbers of novel variants that are uncovered in each exome or genome. The traditional evaluation of genetic variation in Mendelian disease has hinged on several widely accepted criteria for causality.1 Before a variant can be considered potentially causal, it would normally be expected to abolish the synthesis of the encoded protein or to modify, through coding or splicing, amino acid residues of functional significance. Once this initial threshold has been attained, cosegregation with disease through a substantial number of meioses is expected. Functional significance is typically inferred from tight conservation across phyla or through specific biological assays. However, it may be difficult to relate effects in in vitro assays to the biological mechanisms responsible for disease.2 Additional variants meeting these criteria in the same gene would be anticipated in other kindreds with the same phenotype. The case for causality would be further buttressed by the documentation of de novo mutations in the gene in sporadic cases and by the generation of genetically accurate animal models that recapitulate the original phenotype.3Article see p 602 These stringent requirements have been met to date for only a small minority of mutations annotated as definitive or high-confidence in databases of human disease-related genetic variation.4 As the numbers of control sequences available for most genes has increased, it has become clear that many reported mutations in known disease genes are present as rare variants in normal individuals. It is now apparent that the scale and nature of interindividual variation have been substantially underestimated. Indeed, an early insight from exome sequencing has been the lack of trustworthiness of most existing human disease mutation catalogs.5,6 …