Metatranscriptomics improves the laboratory diagnosis of infectious intestinal disease from human diarrhoeal samples

Background Traditional laboratory-based surveillance of gastrointestinal pathogens is time-consuming, which can impact successful outbreak detection. The INTEGRATE study investigated the utility of metagenomic and metatranscriptomic sequencing for the rapid diagnosis of community-associated gastrointestinal infections.

Methods We performed an observational study using stool samples from 1,407 patients with acute gastroenteritis, recruited via general practitioners in the UK. 1,067 stool samples were processed using i) routine clinical methods, ii) a molecular multiplex real-time polymerase chain reaction (PCR) assay, and iii) DNA and RNA sequencing. The relationship between assigned taxonomy, routine clinical diagnostics, and PCR was determined with multivariable linear regression models.

Findings There is a strong, positive relationship between the identification of pathogens in metatranscriptomic reads, and positive results from traditional diagnostics for five out of fifteen pathogens: Campylobacter (p<0·001), Cryptosporidium (p<0·001), Salmonella (p<0·01), Rotavirus (p<0·001) and Sapovirus (p<0·001). Metagenomic sequencing displayed this relationship for two out of fifteen pathogens: Campylobacter (p<0·001) and Salmonella (p<0·01).

Strong positive relationships between metatranscriptomic reads and positive PCR results were observed for six out of fourteen pathogens: Adenovirus, Campylobacter, Cryptosporidium, Norovirus, Rotavirus and Sapovirus (p<0·001). In metagenomic data, the same relationship was observed for four out of fourteen pathogens: Adenovirus (p<0·001), Campylobacter (p<0·001), Salmonella (p<0·05) and Shigella (p<0·01).

A comprehensive transcriptomic profile of Salmonella Enteritidis was recovered from the stool of a patient with a subsequently confirmed Salmonella infection.

Interpretation The metatranscriptomic strategy successfully detected a variety of gastrointestinal pathogens and provided viral and bacterial gene expression profiling directly from stool. We propose that metatranscriptomics could be considered for the future surveillance of gastrointestinal pathogens.