Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore a comprehensive talk by Garyk Brixi discussing Evo 2, a biological foundation model trained on 9.3 trillion DNA base pairs from a curated genomic atlas spanning all domains of life. Learn how this model, available with 7B and 40B parameters, features an unprecedented 1 million token context window with single-nucleotide resolution. Discover how Evo 2 accurately predicts functional impacts of genetic variation without task-specific finetuning, including noncoding pathogenic mutations and clinically significant BRCA1 variants. Understand how mechanistic interpretability analyses reveal Evo 2's autonomous learning of biological features like exon-intron boundaries, transcription factor binding sites, protein structural elements, and prophage genomic regions. See how the model generates mitochondrial, prokaryotic, and eukaryotic sequences at genome scale with improved naturalness and coherence, and how inference-time search enables controllable generation of epigenomic structure. The talk references the fully open Evo 2 model, including parameters, training code, inference code, and the OpenGenome2 dataset, aimed at accelerating exploration and design of biological complexity.
