A stepwise short- and long-read whole genome sequencing strategy resolves previous genetically unresolved inherited retinal disease cases

Stepwise strategy of short-read WGS reanalysis followed by selective long-read WGS to resolve inherited retinal disease cases left without a diagnosis

A significant proportion of rare inherited retinal disease (IRD) cases remain genetically unresolved after short-read whole genome sequencing (WGS), and this research study aimed to increase the diagnostic yield of two cohorts. The first (120 cases) had already been screened by short-read WGS, whereas the second (28 cases) had only undergone exome sequencing (ES) or single-molecule molecular inversion probe (smMIPs) sequencing as pre-screening. Short-read data from 148 cases were (re)analysed with updated allele frequency databases, new variant caller and predictor tools, and updated and extended gene panels to identify causal single nucleotide and structural variants, and long-read WGS was then performed for 20 cases that remained unresolved and had sufficient high-quality DNA. Reanalysis of the 120 previously WGS-screened cases gives a diagnostic yield of 15% (18/120), WGS of the 28 cases without prior WGS an additional yield of 32% (9/28), and long-read WGS two diagnoses among the 20 probands sequenced; the new diagnoses are attributed notably to variants in genes newly associated with IRD or absent from the panels of previous studies, and the Mobile Element Locator Tool identified two pathogenic Alu insertions on short-read data. The authors conclude that periodic reanalysis of existing short-read WGS data is a key component of IRD genetic diagnostics, with selective long-read WGS a valuable additional approach for unresolved cases.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The result to retain is that 15% of cases left undiagnosed after a first short-read WGS are resolved by simple reanalysis, notably because panels, frequency databases and IRD-associated genes have evolved. Long-read sequencing resolves only two cases in 20 in this setting, making it a selective complement rather than a first-line test, and the abstract does not allow the cost per additional diagnosis to be assessed. The two cohorts do not share the same screening history, so 32% and 15% are not comparable with each other; the study was carried out in a research setting and is a preprint.

Analysis by Dr Thibaut Benquey