The team from the Faculty of Computer Science at the University of Białystok is among an international group of scientists who innovative reseach opens new possibilities for analyzing the spread of antibiotic resistance in bacteria using artificial intelligence. The results of this research were published in the journal Frontiers in Genetics.
Reseach on atibiotic resistance in various urban microbiomes was conducted by two collaborating teams from the Jagiellonian University and the University of Białystok. By integrating machine learning, metagenomics and resistance analysis, scientists investigated how resistance genes move between spaces and hospitals.
Rodolfo Brizola Toscan and prof. Paweł Łabaj from the Małopolska Center of Biotechnology, in cooperation with Critical Assessment of Massive Data Analysis (CAMDA) and the Meta SUB project were responsible for data preparation, bioinformatics analyzes and biological analisis.
The team from the University of Białystok includes dr Wojciech Lesińki and mgr Piotr Stomma from the Faculty of Computer Science, dr Balakrishnan Subramanian from the University Computing Center and prof. Witold Rudnicki representing both units used machine learning methods to find key markers differentiating the location of the samples.
The study analysed 143 environmental samples taken from cities and 145 bacterial isolates isolated in hospitals to test their resistance to antibacterial drugs. The data was obtained through cooperation with the international CAMDA (Critical Assessment of Massive Data Analysis) and MetaSUB projects. By comparing resistance markers in bacteria found in hospitals and in urban environments, researchers first evaluated the available tools for detecting resistance. Then, the association of resistance markers with markers of genetic motivity was investigated.
- Our task was to analyze data using machine learning algorithms (i.e. one of the fields of artificial intelligence), developed by a team from the University of Białystok - says Witold Rudnicki, professor at the University of Białystok and director of UCO.
- Using our team's algorithms to identify important traits, we were able to identify key markers of antibiotic resistance. Then, with the help of the random forest algorithm, it was possible to correctly assign samples to the location. Machine learning models helped to better understand the dynamics of AMR (Antimicrobial Resistance) in cities, improving the precision of resistance profiling tools" - adds dr Wojciech Lesiński.
These findings can help bridge the gap between environmental surveillance and nosocomial infections, improving early detection and the development of response strategies.
Properly used, they can give us new tools to fight antibiotic resistance which is becoming one of the most important global threats to public health.
More information about the project and about the research of the environmental microbiome can be found on the website of the Jagiellonian University.
