Novel opportunities for NGS-based one health surveillance of foodborne viruses

Abstract

Foodborne viral infections rank among the top 5 causes of disease, with noroviruses and hepatitis A causing the greatest burden globally. Contamination of foods by infected food handlers or through environmental pollution are the main sources of foodborne illness, with a lesser role for consumption of products from infected animals. Viral partial genomic sequencing has been used for more than two decades to track foodborne outbreaks and whole genome or metagenomics next-generation-sequencing (NGS) are new additions to the toolbox of food microbiology laboratories. We discuss developments in the field of targeted and metagenomic NGS, with an emphasis on application in food virology, the challenges and possible solutions towards future routine application.

Keywords: Food virology; Foodborne virus; Human enteric virus; Metagenomics; Next-generation sequencing; Norovirus.

Fig. 1

Overview of current NGS strategies for virus sequencing in food. Among nucleic acids extracted from virus-contaminated foods, DNA and RNA material from the matrix and bacteria often prevail (blue), and RNA from the contaminating virus (red, green strains) are scarce. Two strategies use specific primers to focus the sequencing power on a viral contaminant previously identified by other means (qRT-PCR). The “metabarcoding” strategy targets regions of the viral genome commonly sequenced for genotyping. If food products are contaminated by several viral strains belonging to different genotypes (red and green strains), PCR products are synthetized for each strain. Deep sequencing of these amplicons results in a mix of reads corresponding to these different strains. Following bioinformatics analyses (mapping or clustering), each read is assigned to one genotype. This approach allows the identification of strains at the genotype level, as well as an estimation of genotype diversity in the sample, for a given virus. The “full genome” strategy uses several sets of primers to amplify overlapping segments spanning the entire viral genome (around 7–8 kb for most enteric viruses). These PCR products are sequenced together using NGS, generating reads that can be assembled into full or partial viral genomes. Depending on the depth and width of coverage, this can allow the identification of the viral strain, including its genotype classification, and comparison with other samples for analysis of transmission pathways, contamination sources, etc.… A third method, metagenomics, uses random primers for cDNA synthesis. With direct deep sequencing of nucleic acid extracted following fast and simple methods, mostly reads from the matrix or bacterial microbiota (blue) are generated, with often only a limited amount of reads corresponding to the virus (red and green), reflecting the contamination level. Identifying the genotype of the viral strain is possible when such reads fall into the typing region of the virus (green). However, if this not the case virus identification may rely only on sequence comparison with databases (red). Additional steps during library preparation (filtration of bacteria, removal of free nucleic acids, exclusion of the matrix RNA using non-random primers for cDNA synthesis, enrichment in viral sequences using hybridization probes) can result in longer viral sequences (red) and possibly full genomes (green) useful for genotyping and studying the transmission pathways. Importantly, this strategy also allows for the potential discovery of new or unexpected viruses (blue), among the vast diversity of nucleic acid sequenced, depending on the stringency of potential selection/enrichment steps