advertisement
advertisement
HANNOVER MESSE 2019, 01 - 05 April
switch to:
Artificial Intelligence

Machine learning detects dangerous salmonella

Artificial intelligence can also be used in medical technology in a practical and useful way. This is shown by a new tool, some of which was developed in Würzburg, with which dangerous pathogens can be discovered faster than before.

14 Jun. 2018
HMI-ID06-002ds_hzi
Machine learning detects dangerous salmonella (Photo: HZI/Manfred Rohde)

No question, life sciences are already well advanced - just think of the decoding of the human genome. But identifying genetic changes in new pathogen strains in an outbreak of a disease will still take a long time. For example, some salmonella can cause food poisoning, while others can cause much more damage throughout the body. A team of researchers, including the Helmholtz Institute for RNA-based infection research in Würzburg, intends to shorten the identification process by means of machine learning. It has therefore developed a software that analyzes what mutations are crucial in the development of a disease. The program was trained to identify important differences between salmonella strains. It then detected nearly 200 genes that influence which disease salmonella causes.

It can also be used to identify new strains of bacteria that pose a potential public health risk. And in a matter of seconds - and not in laborious weeks of painstaking work. The tool can be used on a worldwide basis, is not limited to salmonella and has already passed its first practical test.

Many other projects are currently exploring the potential of machine learning for other areas of medical technology. For example, classification algorithms for breast or lung cancer are already being used in the screening area, reports medica.de . In addition, chemical compounds are evaluated by machine to develop better medications. Also promising is a US-American test, which the Ärzteblatt [Medical Journal] has presented: Six key words and machine learning were enough to distinguish between suicidal and mentally healthy people.

advertisement