The causes of retinal dystrophy and the development of more comprehensive screening approach

Inherited retinal diseases (IRDs) are a group of genetically and phenotypically heterogenous disorders caused by variants in around 280 genes. Additional loci have also been localised to chromosomal regions, though the causative genes remain unknown. Recent improvements in screening technologies have increased the detection of pathogenic variants in IRD. This thesis describes the use of next generation sequencing (second (short-read) and third (long-read) generation sequencing) to find missing or hard to find pathogenic variants in IRD patients.
The first results chapter describes use of whole exome sequencing to screen 24 individuals with syndromic and non-syndromic IRDs. This identified pathogenic variants in known genes in eight cases; CDHR1 (c.1527T>G, p.Y509*), RHO (c.284T>C, p.L95P), PRPF31 (c.797delC, p.S266*), CNGA3 (c.1088T>C, p.L363P), BBS10 (c.728-731delAAGA, p. K243Ifs*15), USH2A (c.252T>G, p.C84W), ABCA4 (c.2588G>C, p.G863A and c.6089G>A, p.R2030Q), and SLC25A46 (c.670A>G, p.T224A). In addition, several candidate variants were highlighted for further investigation.
In the second results chapter, seven patients with late onset macular dystrophy and one with age related macular degeneration were found to carry the same heterozygous ~126 kb deletion encompassing CRX, TPRX1 and SULT2A1. This phenotype has already been documented in patients with heterozygous variants in the gene encoding retinal transcription factor CRX, while there is no known functional or phenotypic link with variants in TPRX1 or SULT2A1. This therefore confirms that CRX haploinsufficiency is pathogenic, a finding that had previously been debated in the ophthalmic literature. The deletion was characterized using a PCR assay followed by cloning and Sanger sequencing or direct Sanger sequencing. Haplotype analysis was done by microsatellite genotyping.
The third results chapter describes use of SMRT PacBio and nanopore long-read sequencing to screen the hard-to-sequence mutation hotspot RPGR-ORF15. Both approaches were effective in reading throughout ORF15 and allowed sequencing indexed pooled samples, and 218 IRD patients were screened, detecting known and new variants. Nanopore sequencing on the smaller Flongle flowcell allowed low-cost optimisation, but pores rapidly blocked, probably due to ORF15 secondary structures. Repeated DNase I washes reopened the pores but required use of the more expensive MinION flowcells. Ultimately, the PacBio sequencer proved simpler to use, cheaper, and more scalable.