USC Launches 7B Parameter METAGENE-1 Model for Pandemic Monitoring and Pathogen Detection

Jan 8, 2025, 04:11 PM

Researchers from the University of Southern California, Prime Intellect, and the Nucleic Acid Observatory have developed METAGENE-1, a 7-billion-parameter autoregressive transformer model designed for global pathogen detection and pandemic monitoring. Trained on over 1.5 trillion DNA and RNA base pairs sourced from human wastewater samples, METAGENE-1 aims to capture the full distribution of genomic information within the human microbiome. This model is designed to aid in public health applications such as pathogen detection and the early identification of emerging health threats. METAGENE-1 has demonstrated state-of-the-art performance in pathogen detection and metagenomic embedding benchmarks, showcasing its potential for biosurveillance and anomaly detection.