# Identification of pathogenic fungi causing ocular infections using full rRNA operon sequencing with Oxford Nanopore Technologies

## TL;DR

This study explores using full rRNA operon sequencing with Oxford Nanopore to identify fungi causing eye infections, offering better accuracy than traditional methods.

## Contribution

The study introduces full rRNA operon sequencing with Oxford Nanopore as a novel diagnostic tool for fungal ocular infections.

## Key findings

- Full rRNA operon sequencing provided comprehensive fungal identification with high confidence for most isolates.
- NGSpeciesID outperformed other pipelines in species-level classification of fungal isolates.
- Oxford Nanopore sequencing revealed limitations in taxonomic resolution for certain fungi like Aspergillus.

## Abstract

Although fungal eye infections are a major cause of visual impairment worldwide, standard clinical laboratory methods remain slow, insensitive, and limited in their taxonomic resolution. Sequencing of the full ribosomal RNA (rRNA) operon provides a comprehensive marker for fungal identification. In this study, twenty fungal isolates associated with ocular infections were obtained from Srinagarind Hospital, Thailand, and characterized using four identification approaches. Initial hospital-based routine identification relied on conventional morphological methods and matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS). To enhance resolution and to develop a comprehensive analytical pipeline, we further employed full rRNA operon sequencing using Oxford Nanopore Technologies (ONT), analyzed through three bioinformatic pipelines: EPI2ME/Minimap2, NGSpeciesID with BLASTn, and internal transcribed spacer (ITS)-based phylogenetic analysis coupled with phylogenetic analysis. All isolates yielded complete operon sequences, thus ensuring comprehensive coverage of the target regions. NGSpeciesID produced high-confidence consensus sequences and species-level classifications for nearly all isolates (except one Candida specimen). Of these, 15 of the 20 isolates showed exhibited concordance with hospital identifications at the genus level (≥97% identity). This approach successfully resolved closely related Aspergillus taxa (i.e., A. terreus, A. luchuensis, A. oryzae), reclassified Curvularia isolates as Bipolaris maydis, and confirmed species-level assignments for Fusarium and Rhodotorula. By contrast, the EPI2ME workflow produced more variable classifications, providing species-level assignments for Aspergillus and Rhodotorula but mixed genus/species profiles for several isolates, including seven isolate assignments unique to this method. ITS-based phylogenetic reconstruction recovered all expected clades, with Curvularia isolates clustering within their genus. However, node support varied substantially, highlighting the limited discriminatory power of ITS alone, which constrains taxonomic resolution to the species-complex level rather than consistently achieving the species-level identification of Aspergillus isolates. Overall, ONT-based full-operon sequencing demonstrates strong potential for fungal diagnostics, its performance depends on bioinformatic pipelines, database quality, and sequencing errors. Species-level resolution is particularly limited in Aspergillus, while incomplete reference datasets hinder the classification of isolates such as Curvularia. To improve reliability and clinical application, it will be essential to expand curated full-length rRNA references, integrate complementary loci, and refine analytical strategies.