Remember meForgot password?
    Log in with Twitter

article imageUsing computers to spot global pathogen spread

By Tim Sandle     Sep 28, 2016 in Science
Edinburgh - The digital age is making inroads into microbiology and epidemiology. Researchers have developed software to help track and to predict pathogenic infections around the world.
The new software is not simply a collection of algorithms used to make predictions; the developers claim that it can ‘learn’ and become more powerful and accurate as time progresses. Specifically it is said that machine learning can predict the strains of bacteria likely to cause food poisoning outbreaks. The software is said to learn the DNA signatures that are associated with pathogen samples that have caused outbreaks of infection in people.
The University of Edinburgh (Roslin Institute) software compares genetic information from bacterial samples isolated from both animals and people, and uses this information to create a map.
To test this out a study was run to examine incidences of Escherichia coli, with a view of aiding public health officials in tracking associated incidences and disease cases. The focus was on strain O157.
This bacterial strain originates from animals, with a high incidence in cows. Cows rarely show signs of illness; however they excrete the pathogen in their feces. Dangers arise when the feces comes into contact with food (such as being used as manure) and thus enters the food chain. Transmission is via the fecal–oral route. Other sources of infection come from raw milk from goats, sheep and cattle, and via the distribution of contaminated leafy green vegetables and undercooked meat.
By gathering genetic information about the bacterium, the software can predict if an E. coli strain came from a cow or a person. This helps to structure interactive maps of potential disease spread, as well as helping to track down the disease origin.
The researcher who led the exploration of the software, Professor David Gally, said in a research briefing: “Our findings indicate that the most dangerous E. coli O157 strains may in fact be very rare in the cattle reservoir, which is reassuring. The study highlights the potential of machine learning approaches for identifying these strains early and prevent outbreaks of this infectious disease.”
It is hoped, after further trials, that the software can be used to track other microbial infections and be used by health authorities around the world.
The software has been reported to the Proceedings of the National Academy of Sciences. The research paper is titled “Support vector machine applied to predict the zoonotic potential of E. coli O157 cattle isolates.” The research was funded by Food Standards Scotland and the Food Standards Agency, both of which are agencies of the Scottish government.
More about Epidemiology, Pathogens, Disease, Infection
More news from
Latest News
Top News