Genetic basis of inherited retinal disease in a UK cohort of over 2900 families

Abstract

Purpose :
Sequence variants in >250 genes have been identified as causing inherited retinal disease. In our molecularly characterised cohort of nearly 3000 families, we sought to quantify proportions of families with causative variants in each gene.

Methods :
We retrospectively searched the database of the retinal genetics service of Moorfields Eye Hospital, London, UK. Molecular diagnoses were sought using a combination of single gene testing, panel testing, whole exome sequencing, and, increasingly, whole genome sequencing. For this study we included those genes listed on the Retinal Information Network online resource (www.retnet.org accessed 14 Nov 2018). We also extracted transcript length for the wild type version of each gene (Ensembl, release 94).

Results : Our search yielded 2928 families in whom a likely causative gene had been found for their phenotype, with 135 disease-causing genes in total. The 20 most prevalent genes accounted for 71.2% of the cohort. These were ABCA4 (20.0%), USH2A (8.9%), RPGR (5.7%), PRPH2 (4.6%), BEST1 (3.9%), RP1 (3.5%), RHO (3.4%), RS1 (3.3%), CHM (2.8%), CRB1 (2.2%), MYO7A (1.8%), PRPF31 (1.7%), OPA1 (1.5%), CNGB3 (1.4%), GUCY2D (1.4%), RPE65 (1.4%), RDH12 (1.3%), CNGA3 (1.1%), and PROM1 (1.1%). 14.5% of families had variants in X-linked genes; 0.9% had variants in mitochondrial genes; 84.6% had variants in autosomal genes (51.4% with exclusively autosomal recessive inheritance; 25.1% in genes in which different variants can show either dominant or recessive inheritance; 8.1% in genes associated only with dominant inheritance). The Spearman coefficient for correlation between number of affected families and normal transcript length was 0.21 (p=0.015).

Conclusions :
Our findings help quantify the burden of inherited retinal disease attributable to each gene in terms of numbers of affected families. Over 70% of families had disease-causing variants in one of only 20 genes. Transcript length, of relevance in developing strategies for gene delivery, correlated significantly with numbers of affected families: larger transcripts tended to have greater numbers of families with disease attributable to variants in those genes, but the correlation was weak.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.