Sequencing based methods for distinguishing viability

Although high-throughput DNA sequencing-based methods have been of great value for determining the composition of microbial communities in various environments, there is the potential for inaccuracies arising from the sequencing of DNA from dead microorganisms. The differentiation of live/dead cell states is particularly important when making decisions with regard to food safety or spoilage, when DNA from cells recently inactivated through food processing/preservation steps can be problematic. Addressing this issue will be of key importance with respect to harnessing the power of DNA sequencing for food microbiology testing. In this study, we compared different sequencing-based methods to assess their relative accuracy with respect to distinguishing between viable and non-viable cells, using a live and heat-inactivated mock community spiked into bovine milk. The methods used were shotgun metagenomics with and without propidium monoazide (PMA) treatment, RNA-based 16S rRNA sequencing and metatranscriptomics. The results showed that methods were generally accurate, though significant differences were found depending on the library types and sequencing technologies. Different molecular targets were the basis for variations in the results generated using different library types, while dissimilarities in the results from Oxford Nanopore Technologies- and Illumina-based sequencing likely reflect a combination of different sequencing depths, error rates and bioinformatics pipelines. Although PMA was successfully applied in this study, further optimisation is required before it can be applied in a more universal context for complex microbiomes. Overall, these methods show promise and represent another important step towards the ultimate establishment of approaches that can be applied to accurately identify live microorganisms in milk and other food niches.