Metagenomics to Detect and Characterize Viruses in Food Samples at Genome Level? Lessons Learnt from a Norovirus Study

In this proof-of-concept study on food contaminated with norovirus, we investigated the feasibility of metagenomics as a new method to obtain the whole genome sequence of the virus and perform strain level characterization but also relate to human cases in order to resolve foodborne outbreaks. We tested several preparation methods to determine if a more open sequencing approach, i.e., shotgun metagenomics, or a more targeted approach, including hybrid capture, was the most appropriate. The genetic material was sequenced using Oxford Nanopore technologies with or without adaptive sampling, and the data were analyzed with an in-house bioinformatics workflow. We showed that a viral genome sequence could be obtained for phylogenetic analysis with shotgun metagenomics if the contamination load was sufficiently high or after hybrid capture for lower contamination. Relatedness to human cases goes well beyond the results obtained with the current qPCR methods. This workflow was also tested on a publicly available dataset of food spiked with norovirus and hepatitis A virus. This allowed us to prove that we could detect even fewer genome copies and two viruses present in a sample using shotgun metagenomics. We share the lessons learnt on the satisfactory and unsatisfactory results in an attempt to advance the field.

Automated summary

This proof-of-concept study explores the feasibility of using metagenomics to obtain whole genome sequences of norovirus from contaminated food and perform strain level characterization, as well as relate to human cases. The study compares different sequencing methods and demonstrates that shotgun metagenomics is effective for phylogenetic analysis while hybrid capture is suitable for lower contamination. The study also shows the potential of shotgun metagenomics to detect multiple viruses in food samples.