Abstract
Purpose :
Genome-wide association studies of complex eye disease have found many disease susceptibility loci, but the causal genes and variants at these loci generally remain unknown. One way to help identify causal variants is to study RNA editing, a post-transcriptional modification to an RNA nucleotide. The most common modification is the editing of an adenosine to inosine (A-to-I) thus fooling the ribosome to think it is translating a guanine. We hypothesize that RNA editing is not uniform throughout the eye and in fact may be contributory to localized disease in macula.
Methods :
Human eyes (n=7) with normal maculas were collected within 6 hours of donor death. cDNA libraries made from the macula and periphery of the retina (R) and of the RPE/choroid/sclera (RPCS) of each eye were sequenced on an Illumina HiSeq. Reditools software was used to predict significant editing sites using Fisher Exact Test and FDR correction (p < 0.05). Known single nucleotide polymorphisms from dbSNP, 1000Genomes, and SeattleSeq were removed to reduce false positives. After a preliminary list of predicted A-to-I editing sites was produced, BLAT was used to confirm and filter the original GSNAP alignment. After processing, RNA editing sites were annotated and quantified between locations (macula versus periphery) and tissues (R versus RPCS).
Results :
The number of significant editing sites in the macular R and RPCS and peripheral R and RCS were 1,147, 501, 997, and 501, respectively. For each tissue-location, 8-16% of the predicted editing sites were known from the DARNED database, 22-32% of the editing sites were located within protein coding genes, and 70% of the editing sites occur within noncoding RNAs. In addition, 30 editing sites were found within age-related macular degeneration GWAS loci with 5 of them in 3’ UTR of protein coding genes.
Conclusions :
Our results support our hypothesis that RNA editing is not uniform throughout the R and RPCS. The increased number of editing sites in R compared to RPCS is striking (p = 8.8x10-9). SNPs within these editing sites, especially if located within protein coding regions, could be contributory to disease and deserve more attention when we try to understand the genetic architecture of disease. More attention needs to be directed to the influence of RNA editing in noncoding RNAs.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.