Garyk Brixi
147 posts
@garykbrixi
Building discovery engines for biology. @Stanford · @ArcInstitute
Many of the most important drug classes in modern history were nearly abandoned by their financial backers. If we can solve the structural risk-aversion that almost prevented these drugs from getting to patients, then we can dramatically accelerate medical progress.
@arcinstitute @pdhsu I tested whether Evo2's embedding space could work as a 'semantic BLAST' — retrieval by function instead of sequence alignment. 25 genes, 7B model, 475 windows. Here's what I found: The top of the raw similarity ranking is repeat-driven — L1, Alu, SVA(jumping genes) (1/4)
Evo 2 is out in Nature today, showing that genome language models can predict and design across the full complexity of life, from phages to eukaryotes. A few surprises from the project, including how ignoring trillions of nucleotides was key to getting a good model. 🧵
Evo 2 is out in Nature today, showing that genome language models can predict and design across the full complexity of life, from phages to eukaryotes. A few surprises from the project, including how ignoring trillions of nucleotides was key to getting a good model. 🧵
Evo 2 is out in Nature today, showing that genome language models can predict and design across the full complexity of life, from phages to eukaryotes. A few surprises from the project, including how ignoring trillions of nucleotides was key to getting a good model. 🧵
Thrilled to see the Evo2 paper led by @BrianHie, @pdhsu & team out in @Nature! We helped bring long (20kb!) AI-designed DNA to life in cells. Seeing experiments match the designs was wild-biological abstraction is starting to feel real. Join us → pinglay-lab.com
Evo 2 is out in Nature today, showing that genome language models can predict and design across the full complexity of life, from phages to eukaryotes. A few surprises from the project, including how ignoring trillions of nucleotides was key to getting a good model. 🧵
Thrilled to see the Evo2 paper led by @BrianHie, @pdhsu & team out in @Nature! We helped bring long (20kb!) AI-designed DNA to life in cells. Seeing experiments match the designs was wild-biological abstraction is starting to feel real. Join us → pinglay-lab.com
I wanted to post some fun GIFs from the Evo 2 inference time compute story! In the preprint, we show that we can use Evo 2 to propose sequences and guide generation with Enformer/Borzoi to control aspects of epigenomic architecture. A very simple algorithm that worked well!
