LLMs Coming For A DNA Sequence Near You

An illustration of two translucent blue hands knitting a DNA double helix of yellow, green, and red base pairs from three colors of yarn. Text in white to the left of the hands reads: "Evo 2 doesn't just copy existing DNA -- it creates truly new sequences not found in nature that scientists can test for useful properties."

While tools like CRISPR have blown the field of genome hacking wide open, being able to predict what will happen when you tinker with the code underlying the living things on our planet is still tricky. Researchers at Stanford hope their new Evo 2 DNA generative AI tool can help.

Trained on a dataset of over 100,000 organisms from bacteria to humans, the system can quickly determine what mutations contribute to certain diseases and what mutations are mostly harmless. An “area we are hopeful about is using Evo 2 for designing new genetic sequences with specific functions of interest.”

To that end, the system can also generate gene sequences from a starting prompt like any other LLM as well as cross-reference the results to see if the sequence already occurs in nature to aid in predicting what the sequence might do in real life. These synthetic sequences can then be made using CRISPR or similar techniques in the lab for testing. While the prospect of building our own Moya is exciting, we do wonder what possible negative consequences could come from this technology, despite the hand-wavy mention of not training the model on viruses to “to prevent Evo 2 from being used to create new or more dangerous diseases.”

We’ve got you covered if you need to get your own biohacking space setup for DNA gels or if you want to find out more about powering living computers using electricity. If you’re more curious about other interesting uses for machine learning, how about a dolphin translator or discovering better battery materials?

13 thoughts on “LLMs Coming For A DNA Sequence Near You

    1. Hallucinations of the Deep Dream sort likely won’t be viable, it’s the AI “landscape pictures” and “background scenes” that seem to be everywhere that will be the problem.

      You know, the ones with winding mountain roads that run in a circle, streams flowing uphill, and people in a family scene with seven fingers on one hand and three fingers each on the other two hands from other wrist that will give us problems.

    1. I have no idea about that either.

      The only thing I could find on a search was the “Moya gene” in the “Moyamoya disease (which) is a chronic and progressive condition of the arteries in the brain. People with moyamoya disease have narrowing of these blood vessels that leads to blockages and can eventually cause ischemic stroke, hemorrhagic stroke, and seizures.” which I am certain was not the Moya intended.

      1. Think there’s a Spanish footballer called Moya?

        Let’s hope the LLM was trained on FIFA live plays not medical journals, and doesn’t give you Moyamoya disease!

        What could possibly go wrong?!

  1. been waiting for an LLM that I can run locally that can understand DNA. then I could load my data from my full genome and have a conversation with my DNA, family lineage, etc. Combine with my local agent’s knowledge of my medical records, it might be a better opinion than any doctor or cloud service might give.

Leave a Reply to SammyCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.