Abstract
Purpose :
While retinitis pigmentosa (RP) is genetically heterogenous disease, mutations in EYS gene are the most frequent cause in Japanese patients with RP accounting for 30 - 40% of genetically diagnosed cases. However, pathogenicity of some variants is controversial due to lack of definite criteria. In this study, we examined the details of all EYS variants in Japanese patients with RP.
Methods :
Variants screening was done in 435 RP patients using next generation sequencing. We examined minor allele frequency (MAF) in 3 databases (EXaC, 1000genome, and HGVD) and mutation types (splice sites, frameshift deletion/insertion, nonframeshift deletion/insertion, stopgain, stoploss, missense, and synonymous) of each variant. For missense variants, we evaluated the pathogenicity using 5 in-silico functional prediction tools (SIFT, Polyphen2, LRT, Mutation Tastar, and Mutation Assessor) and 2 in-silico conservation algorithms (GERP++ and phyloP). We defined “in-silico prediction score” as the number of the positive result (pathogenic or likely/possibly pathogenic) among the five functional prediction tools.
Results :
Seventy seven EYS variants were detected (2 splice sites, 2 nonframeshift deletion, 59 missense, 10 frameshift, and 4 stopgain variants). All of the frameshift and stopgain variants had extremely low MAF in all MAF databases (12 of 14 variants had 0 in all the databases). Among the 59 missense variants detected, 7 variants scored ≥ 3 in in-silico prediction score, and all variants other than one (c.2528G>A:p.G843E) had low MAF in all thedatabases. Among the 52 variants with < 3 in-silico prediction score, which namely are regarded as non-pathogenic variant in some of the previous reports, 21 variants had very low MAF (< 0.001 in all the databases), and 10 of them had MAF of 0 in all MAF database.
Conclusions :
We showed that some variants judged as non-pathogenic based on in-silico programs have extremely low MAF. Considerable number of variants with very low MAF found in RP patients suggest these are actually disease causing mutations.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.