SEARCH

SEARCH BY CITATION

CHARGE syndrome is an autosomal dominant complex malformation disorder caused by mutations in CHD7. While ∼90% of the mutations in this gene are easily recognized as pathogenic, ∼8% of patients carry a missense substitution that would be called an unclassified variant at first observation. In the absence of a functional assay for CHD7 missense substitutions, Bergman et al. (Hum Mutat 33:1251–1260, 2012) have developed a computational approach to classify missense substitutions in this gene.

Their task was confounded by two limitations: (1) there were no gold standard “clearly pathogenic” and “clearly neutral” missense substitutions to use in algorithm development, and (2) the number of known missense substitutions and amount of data available for each substitution are not sufficient to quantitatively calibrate the analytic methods that the authors sought to integrate into a single evaluation. To overcome these challenges, the authors started with the 145 CHD7 missense substitutions reported in the literature, dbSNP, or recorded by Dutch or Danish testing labs. Using clinical criteria, they developed a subset of 12 very likely benign and 9 very likely pathogenic substitutions, which were used to evaluate the performance of the multiple alignment-based algorithms SIFT, PolyPhen-2, and Align-GVGD, plus the completely structural algorithm FOLDX. Output from the most successful algorithms was integrated with observational data including de novo occurrence and co-occurrence with clearly pathogenic mutations.

The final classification system recapitulated the initial 21 confidently classified substitutions exactly. Overall, 59 of 145 substitutions were classified as probably benign, 46 as uncertain, and 40 as probably pathogenic. The system should be a valuable tool for analysis of clinically observed CHD7 substitutions. Moreover, this classification method framework should prove to be a useful template for other genes and syndromes that face the problem of unclassified missense substitutions with insufficient data to perform the necessary quantitative methods calibrations.