Hmm, the CCD code option disappeared from the #AlphaFold3 server...

An update to #AlphaFold3 server. Now, you can add any ligand from the PDB using a CCD Code (up to 2024_10_28). Here is an MPro #SARS-CoV2 inhibitor from PDB 8u9h. The crystal structure is grey, and the #AF3 model is colored by pLDDT.

Learn how to predict mono-, multi-mers and conformations with the ColabFold-AlphaFold2 protocol now out in @NatureProtocols. Right in time for the Nobel Prize! It covers workflows for beginners and advanced concepts.
🚀https://rdcu.be/dWViQ
📄https://www.nature.com/articles/s41596-024-01060-5 🧬🧶

Outer product mean should go from MSA to pair representation.

David Baker chats with Rosetta folks

Nobel Laureates in Chemistry David Baker, Demis Hassabis, and John Jumper

David Baker visits all the Rosetta folks after his Nobel lecture. And photo of all three Chemistry laureates - David Baker, Demis Hassabis, and John Jumper. An honor to be here in Stockholm to see this.

We updated our BindCraft preprint with lots of new exciting results! We release all our binder sequences and models, include more in silico analysis, novel design targets, and present AAV retargeting to specific cell types using de novo binders!
www.biorxiv.org/content/10.1...

Here is the performance of AlphaFold3-server vs. ColabFold-baseline on protein domains. AlphaFold3 is better for the majority of targets. I suspect that the weaker performance of AlphaFold3 in a couple of cases is due to using a reduced database for MSA construction.

Here is the plot for the protein complex predictions. AlphaFold3 server is an improvement over AlphaFold2 baseline represented by ColabFold. #CASP16

#CASP16 program is posted (so you can guess the "winners". Congratulations to Yang and Kihara who seems to have done well in RNA+Proteins. Also congratulations to AF3-server (i.e. me) who was selected to talk (i.e. most people did worse than the server). predictioncenter.org/casp16/doc/C... .