Redefining Mutational Spectra via Updated Locus-specific Databases
Johan T. den Dunnen · Human Mutation · 2014
Variants in the USH2A gene are responsible for Usher syndrome type II (USH2), a group of autosomal recessive disorders that affect both hearing and vision. The USH2A gene spans 790 kb, has 72 exons and codes for two protein isoforms. The full-length isoform is a 580 kDa transmembrane protein (5,202 amino acids) with a large extracellular domain. The short 170 kDa isoform is translated from only the first 21 exons and is regarded as an extracellular protein. Numerous molecular studies have identified variants in both isoforms (USHBases, http://www.LOVD.nl/USH2A), showing USH2A is responsible for 75–82% of USH2 cases. In this issue, based on an extensively genotyped new cohort of 152 patients, Baux et al. (Hum Mutat 35: 1179–1186, 2014) give a summary of all variants reported to date. All data are presented in a public database including, per variant, the classification of the experts regarding its potential relation to disease (“pathogenicity”). Analysis of the total data set revealed several features of the variant spectrum: while some regions of the protein tolerate amino acid substitutions, other regions are hotspots for deleterious variants. Where possible, all variants identified per patient are reported and such data can be used to look at potential additive effects of variants, including the influence of polymorphic variants. The authors end with an interesting experiment: they retrieved all 466 USH2A missense variants from the Exome Variant Server (EVS) and tried, with the help of bioinformatic prediction tools, to classify these based on the knowledge gained from the variants found in USH2 patients. Though the EVS dataset is considered representative of the general population, an unexpectedly high number of 136 EVS variants came out as putatively deleterious. The authors were clearly surprised by this finding and leave the reader in doubt. Do many individuals contain a deleterious USH2A variant on one allele, or is the predicted classification wrong? The exercise, analyzing data from a resource like the EVS, seems worthwhile to be repeated by others. Test the data yourself-are you able to discriminate variants that are deleterious from those that are not?