Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Paper Club

RFdiffusion3, an atomic-level protein and nucleic acid design AI, published in Science

An international team from the University of Washington, the University of Oxford, Nvidia and other institutions has published the detailed design of RFdiffusion3 in the journal Science, according to IT之家. The diffusion-based system represents a single amino acid with up to 14 atoms and designs new structures by directly adjusting each atom's 3D coordinates, satisfying constraints such as the positions of reacting atoms, hydrogen-bond conditions and permitted atom distribution inside the protein. Comparing designs against 41 existing enzyme active sites, the team reported RFdiffusion3 outperformed RFdiffusion2 in 37 cases and generated structures about ten times faster; a paraoxon-degrading enzyme and a peptide-cleaving enzyme it designed carried out the intended chemistry in real experiments. The report adds that the team open-sourced RFdiffusion3 on December 3, 2025, and that the paper's corresponding author is David Baker, a 2024 Nobel laureate in chemistry. Why it matters: The report says AI protein research has moved from predicting existing structures to designing new proteins with target structures and functions, and that enzymes designed by RFdiffusion3 achieved their intended reactivity in real experiments. What do you think?

Open the headline: source, AI briefing and more →

Comments

…