A team of researchers has found a new way to detect dangerous strains of bacteria, potentially preventing outbreaks of food poisoning. The team developed a method that utilizes machine learning and tested it with isolates of Escherichia coli strains. The details are in a paper that was just published in the journal Proceedings of the National Academy of Sciences.
Most strains of Escherichia coli are harmless and naturally found in the human body. There are pathogenic strains, however, and they are a rising health concern. The E.coli O157 strain in particular is known for causing food poisoning outbreaks and serious illness. Not all forms of E.coli O157 are harmful and the bacteria are commonly found in the guts of cattle. Detection is difficult since infected cows show no signs of disease, even when carrying strains that are dangerous to people.
A research team from the University of Edinburgh’s Roslin Institute used computer software to analyze genetic information from both pathogenic and harmless forms of E.coli O157. The team utilized machine learning, a form of artificial intelligence that allows computers to learn and analyze patterns. With this method, the researchers “trained” the computer to discriminate between human and cattle E.coli strains. The computer was then able to use genetic data to predict if a particular bacterial isolate would be pathogenic to humans.
The researchers found that less than 10% of cattle E.coli O157 isolates were pathogenic. The computer was able to accurately determine whether or not a bacterial strain was harmless based on genetic lineages and patterns. The team’s method could be used with other types of bacteria, including Salmonella.
The team’s findings show promise for the development of new, faster methods of detecting pathogenic bacteria. If researchers can quickly identify cattle herds carrying dangerous strains of E.coli, that particular herd can be treated or isolated before an outbreak occurs.
Nadejda Lupolova et al. Support vector machine applied to predict the zoonotic potential of O157 cattle isolates. Proceedings of the National Academy of Sciences (2016).