Predicting the potential of E. coli transmission from cattle to humans
Computers learn to recognise strains of E. coli present in cattle that are a threat to human health.
Certain strains of Escherichia coli that are normal harmless residents of the gastrointestinal tract of cattle can cause serious human disease if we ingest food or water contaminated with them.
Our recent work has applied a machine learning approach to train computers to ‘recognise’ the subset of strains present in cattle that are a threat to human health.
This finding has important implications for targeting of control strategies in herds and can be used to identify the animal sources from which zoonotic pathogens originate.
For further information please read our news story.
Lupolova N., Dallman TJ, Matthews L, Bono JL & Gally DL (2016). Support Vector Machine applied to predict the zoonotic potential of E. coli O157 cattle isolates. Proc Natl Acad Sci USA, 10.1073/pnas.1606567113.