Culture-free, strain-level characterisation of foodborne pathogens using targeted nanopore metagenomics

Rapid detection and characterisation of foodborne pathogens is essential in outbreak investigation and food manufacturing quality control monitoring, where the source of the microbial contamination needs to be identified to protect the food supply chain and safeguard public health. Traditionally, food samples suspected of causing illness are subjected to culture isolation and subculturing techniques; however, these methods are time-consuming and can result in delays. Even with the development of modern approaches, such as metagenomic sequencing, culture enrichment of the sample has remained a prerequisite, despite the potential to introduce bias from growth competition. Due to these limitations, in 60% of foodborne outbreaks reported in the EU, the contaminated food source responsible cannot be determined with certainty¹.

Nanopore sequencing offers real-time data streaming, which provides immediate access to results, and underpins adaptive sampling — an on-device enrichment method unique to Oxford Nanopore, providing the ability to direct sequencing capacity toward sequences of interest in real time. This means that enrichment for potential pathogens in food sources could be done without the time-consuming steps associated with culture-dependent methods.

Experimental setup

To evaluate whether the need for culture enrichment can be circumvented, Buytaers et al. explored the use of both untargeted nanopore metagenomics and targeted nanopore metagenomics with adaptive sampling in identifying a foodborne pathogen that was previously linked to an outbreak². The experiment was performed in two ways: first, with Staphylococcus aureus DNA — from the strain that had been associated with the outbreak — mixed with potato DNA, and the second, with samples collected from mashed potatoes spiked with the living S. aureus strain.

In the first part of the experiment, the team sequenced the mixed DNA multiple times on a GridION device from Oxford Nanopore Technologies using different techniques: twice via metagenomic sequencing, and three times with adaptive sampling that either depleted potato DNA, enriched for S. aureus DNA, or enriched for DNA of various foodborne pathogens. The aim was firstly to see if the different techniques could detect the pathogen and, secondly, to see if they could characterise it, for example, to uncover its potential to produce toxins.

Pathogen detection and characterisation

Both the untargeted and targeted nanopore metagenomic sequencing approaches successfully identified S. aureus in the sample, indicating that both techniques are suitable for pathogen detection of contaminated food sources. Exploring the capacity of the techniques to identify previously determined, strain-specific enterotoxin genes³, the team found that the use of adaptive sampling enhanced this identification. The adaptive sampling approaches that either depleted potato DNA or enriched S. aureus DNA detected all enterotoxin genes, and the adaptive sampling approach that enriched for foodborne pathogens detected 85% (11/13) of the genes, compared with the untargeted approach that identified 62% (8/13) of them in one replicate and 85% (11/13) in another.

Real-world application

The team observed similar results for the samples collected from mashed potatoes, suggesting that nanopore metagenomics is applicable to real contaminated food sources. When using adaptive sampling to enrich for S. aureus DNA, the team were able to fully characterise the strain and correctly place it within the outbreak cluster on a phylogenetic tree.

‘This method shows great potential for strain-level analysis of foodborne outbreaks without the need for culture enrichment, thereby enabling faster investigations and facilitating precise pathogen characterization’².

Conclusions and recommendations

Overall, Buytaers et al. reported that ‘adaptive sampling outperformed the shotgun sequencing’, and that for real-world application, they noted that using targeted nanopore metagenomics with adaptive sampling based on a database of the suspected pathogen would generate the fastest results without the limitations of culture. If the suspected pathogen is unknown, the team suggests using adaptive sampling with a broader database of foodborne pathogens to detect key genes and virulence factors.

The team concluded that ‘the integration of adaptive sampling with metagenomics presents a valuable strategy for more efficient and targeted analysis of microbial communities in foodborne outbreaks, contributing to improved food safety and public health’.