Abstract
BACKGROUND: Molecular diagnostics such as pathogen-directed PCRs have transformed testing for ocular infections since the late 1990s. Although these assays remain important diagnostic tools for samples with low biomass, the lack of diagnostic range motivates alternative molecular approaches for ocular infections. The aim of this study was to determine the performance of a high-throughput RNA sequencing approach, RNA-seq, to detect infectious agents in ocular samples from patients with presumed ocular infections. METHODS: We compared the performance of RNA-seq to pathogen-directed PCRs using remnant nucleic acids from 41 aqueous or vitreous samples of patients with presumed ocular infections. Pathogen-directed PCRs were performed at the CLIA-certified Stanford Clinical Virology Laboratory. RNA-seq was performed in a masked manner at the Proctor Foundation at the University of California San Francisco. Percent positive and negative agreement between the two testing approaches were calculated. Discordant results were subjected to orthogonal testing. RESULTS: The positive percent agreement between RNA-seq and pathogen-directed PCRs was 100% (95% confidence interval (CI): 78.5%-100%). The negative percent agreement was 92.6% (95% CI: 76.6%-97.9%). RNA-seq identified pathogens not on the differential diagnosis for 9.7% (4/41) of the samples. Two pathogens solely identified with RNA-seq were confirmed with orthogonal testing. CONCLUSIONS: RNA-seq can accurately identify common and rare pathogens in aqueous and vitreous samples of patients with presumed ocular infections. Such an unbiased approach to testing has the potential to improve diagnostics although practical clinical utility warrants additional studies.