From reads to reality : optimizing the performance of long-read nanopore shotgun metagenomics for assembly-free microbial detection

Summarise with VIRAT
Authors: Alexander Van Uffelen
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Shotgun metagenomics theoretically enables sequencing of all DNA in a sample, potentially providing an unbiased view of its microbial composition without the need for prior knowledge or organism isolation. Combined with long-read sequencing technologies like Oxford Nanopore Technologies, it offers a highly promising approach for universal microbial identification. This could make long-read shotgun metagenomics a potential game-changer for public health. A central step in microbial identification from shotgun metagenomic data is taxonomic classification, which identifies the present organisms in an assembly-free matter. However, taxonomic classification of long-read shotgun metagenomic data is mainly performed in research settings. This PhD evaluates the reliability of nanopore sequencing for shotgun metagenomic microbial identification in public health and examines the challenges of extending its use beyond research. In the first step, we built a modular framework to compare the performance of different taxonomic classifiers. Classifiers adhering to a set of requirements, which promote open use, were benchmarked using real nanopore sequenced defined mock communities. The benchmarking provides insight into both the strengths and limitations of each classifier type. The framework was designed to be modular, as few classifiers existed specifically made for nanopore reads, and new developments in this area are expected. Because database choice and other factors also influence classification results, the modular design allows evaluation of much more than just the classifiers themselves. In the second step, we demonstrated the feasibility of nanopore sequencing in a real-world public health application. In a case study where traditional methods such as PCR shows limitation to distinguish closely related species, long-read shotgun metagenomics successfully identified the species present in the sample. The workflow combined in silico development with in vitro validation, and showed that classifier optimization and database curation are crucial for reliable taxonomic classification. In the third step, we expanded the findings from the aforementioned case study to a more broad setting. Here, we examined how different factors such as read filtering and sequencing yield affects classification performance and limit of detection. The final outcome is a publicly available tool optimized for general use, which outperforms other non-optimized classifiers commonly used in research. These three steps show nanopore-based shotgun metagenomics is a valuable asset for microbial identification in public health. The developed framework and optimized tool also provide a strong foundation for further research.

Fields of Study & Subjects

Genomics and Phylogenetic Studies Bacterial Identification and Susceptibility Testing Environmental DNA in Biodiversity Studies

Publication details

Journal / source
Ghent University Academic Bibliography (Ghent University)
Publication date
1 January 2026
Language
EN
Document type
Thesis