News: New AI Breakthrough Can Model And Design Genetic Code Across All Domains of Life.
Published 12:00 AM EST, Weds March 12, 2025
The model, called Evo 2, can identify patterns in gene sequences across disparate organisms that experimental researchers would typically need years to uncover. In addition to identifying disease-causing mutations in human genes, Evo 2 can design new genomes that are as long as the genomes of simple bacteria.
A team of researchers from UC Berkeley, Arc Institute, UCSF, Stanford University, and NVIDIA has developed Evo 2, the largest AI model in biology to date. Trained on DNA from over 100,000 species, Evo 2 identifies patterns in gene sequences that would take years for experimental researchers to uncover. The model is capable of detecting disease-causing mutations in human genes and can even design genomes as long as those of simple bacteria. This marks a significant advancement in biomolecular sciences, leveraging machine learning to accelerate discoveries in genetics and synthetic biology.
Building upon its predecessor, Evo 1, which was limited to single-cell genomes, Evo 2 has been trained on over 9.3 trillion nucleotides from 128,000 genomes, including bacteria, archaea, phages, humans, plants, and other multicellular eukaryotes. By incorporating metagenomic data and a broad representation of life, Evo 2 has achieved a generalist understanding of the tree of life. This makes it a versatile tool for applications ranging from genetic research and evolutionary biology to the creation of artificial life.
Patrick Hsu, UC Berkeley assistant professor and co-senior author, describes Evo 2 as a milestone in "generative biology," where AI models can now "read, write, and think in the language of nucleotides." The model’s ability to analyze vast amounts of genetic information opens doors for groundbreaking applications, including the design of synthetic genomes and better predictions of genetic disorders. Evo 2 sets a new precedent for using AI in life sciences, providing a foundation for further advancements in biotechnology and medicine.
For the cannabis industry, Evo 2 holds immense potential in genomic research and breeding innovation. By analyzing the complex genetics of cannabis strains, the model could help identify genes linked to desirable traits such as cannabinoid profiles, terpene production, and resistance to pests or environmental stress. This could significantly accelerate breeding programs, allowing for the precise engineering of strains tailored for medical, recreational, and industrial use. Additionally, Evo 2’s ability to predict mutations could enhance quality control in cannabis cultivation, ensuring genetic stability and consistency across large-scale production. As AI-driven biology continues to evolve, tools like Evo 2 could revolutionize cannabis genetics, unlocking new possibilities for plant science and commercial applications.
Source: Berkeley Engineering
