Hebrew University scientists unveil AI tool identifying soil bacteria.

Scientists at the Hebrew University of Jerusalem have unveiled an artificial‑intelligence system that can sort through DNA collected from soil and pinpoint the bacteria hidden inside, a breakthrough announced on Tuesday at a conference in Jerusalem.

Soil teems with microbes that drive nutrient cycles, support plant growth and influence climate, yet most of their genetic material remains unclassified. Traditional methods rely on matching DNA fragments to reference libraries, which cover only a fraction of known species. The new AI tool, called SoilSense, was trained on millions of microbial genomes and can recognize patterns that elude human analysts, boosting identification rates by roughly 30 percent in pilot tests.

The research team, working with partners in the United States and Germany, fed SoilSense DNA data from 150 soil samples collected across agricultural fields, forests and deserts. The system flagged dozens of previously unknown bacterial strains and correctly assigned over 90 percent of known organisms, according to the study published in *Nature Biotechnology*. “We can now see a much richer picture of the underground ecosystem,” said Dr. Maya Levi, lead author of the paper. Environmental groups have welcomed the advance, saying it could help monitor soil health and guide sustainable farming practices.

The developers plan to release SoilSense as an open‑source platform later this year, allowing laboratories worldwide to apply the technology to their own samples. They also aim to expand the training set with data from marine sediments and permafrost, hoping to uncover more hidden microbes. If successful, the tool could accelerate research on antibiotic discovery, carbon sequestration and the resilience of ecosystems under climate change.