New AI microbiome tool offers breakthroughs in forensics and epidemiology
Microorganisms, which include tiny bacteria that thrive in different environments, can serve as markers for an individual's recent whereabouts. However, until now, linking a microbiome sample to its precise geographical source was a significant challenge. The mGPS system utilizes advanced AI to map these microbial "fingerprints," allowing it to track samples back to specific locations, such as bodies of water, urban centers, or even distinct transit stations.
"Unlike human DNA, the human microbiome shifts constantly based on environmental contact. By mapping these microbial changes, we can trace the spread of disease, pinpoint infection sources, and even uncover clues for forensic investigations," explained Eran Elhaik, a biology researcher at Lund University and the lead author of the study, which was published in 'Genome Biology and Evolution'.
The study's data show that microbial communities, much like human populations, carry distinct geographical markers. Some bacterial strains are found only in specific locations, creating a unique microbial signature. The team's AI model has been trained to recognize these unique bacterial compositions and link them back to specific geographical coordinates.
Elhaik's team analyzed large microbiome datasets, including samples from urban, soil, and marine environments, and used these to train their model. They reported that mGPS achieved a 92% accuracy rate in determining the source city of city-derived samples. Remarkably, in Hong Kong, the tool identified the precise underground station of a sample with 82% accuracy. In New York City, it could differentiate microbiome samples from a kiosk and nearby handrails just one meter apart.
This capability stems from the vast data compiled by the team, which included 4,135 samples from public transit in 53 cities, 237 soil samples from 18 countries, and 131 marine samples from nine bodies of water. Elhaik sees this as only the beginning: as more microbiome data becomes available, he envisions an expansion in forensic applications that could provide an even more detailed microbial map of cities.
"We have only scratched the surface of understanding how microorganisms reflect their environments. Our next goal is to map microbiomes across entire cities, a project that could further aid forensic science and deepen our understanding of the microbial ecosystems around us," said Elhaik.
Research Report:Microbiome Geographic Population Structure (mGPS) Detects Fine-Scale Geography
Related Links
Lund University
Space Technology News - Applications and Research
