Post

I hear so much about the growing power of predictive biology, especially in drug discovery and "co-folding". I want to see people put their GPU budget where their mouth is!

CYP3A4 metabolizes something like half of the drugs on the market and it is the biggest single source of drug-drug interactions. It would be huge to be able to predict which molecules inhibit CYP3A4 and to modify them to remove this liability!

So if you think you can do structure prediction (or improve structure prediction with MD, etc), you should look at the CYP challenge that we at @openadmet.bsky.social are running right now: huggingface.co/spaces/opena...

I wanted to highlight the structure prediction track because it features 20 novel CryoEM structures of CYP3A4 bound to ligands, collected by Anna Karen Orta!

You might get some useful intel from Karen's recent preprint: www.biorxiv.org/content/10.6...

If you want a head start, go look at what worked in our PXR challenge for structure prediction (do a lot of AI prediction, score with some physics is one strategy!)