Abstract
BACKGROUND: Large databases permit quantitative description of genes in terms of intolerance to loss of function ('haploinsufficiency') and prevalence of missense variants. We explored these parameters in inherited retinal disease (IRD) genes. METHODS: IRD genes (from the 'RetNet' resource) were classified by probability of loss of function intolerance (pLI) using online Genome Aggregation Database (gnomAD) and DatabasE of genomiC varIation and Phenotype in Humans using Ensembl Resources (DECIPHER) databases. Genes were identified having pLI ≥0.9 together with one or both of the following: upper bound of CI <0.35 for observed to expected (o/e) ratio of loss of function variants in the gnomAD resource; haploinsufficiency score <10 in the DECIPHER resource. IRD genes in which missense variants appeared under-represented or over-represented (Z score for o/e ratio of <-2.99 or >2.99, respectively) were also identified. The genes were evaluated in the gene ontology Protein Analysis THrough Evolutionary Relationships (PANTHER) resource. RESULTS: Of 280 analysed genes, 39 (13.9%) were predicted loss of function intolerant. A greater proportion of X-linked than autosomal IRD genes fulfilled these criteria, as expected. Most autosomal genes were associated with dominant disease. PANTHER analysis showed >100 fold enrichment of spliceosome tri-snRNP complex assembly. Most encoded proteins were longer than the median length in the UniProt database. Fourteen genes (11 of which were in the 'haploinsufficient' group) showed under-representation of missense variants. Six genes (SAMD11, ALMS1, WFS1, RP1L1, KCNV2, ADAMTS18) showed over-representation of missense variants. CONCLUSION: A minority of IRD-associated genes appear to be 'haploinsufficient'. Over-representation of spliceosome pathways was observed. When interpreting genetic tests, variants found in genes with over-representation of missense variants should be interpreted with caution.