DeepMind presented remarkably accurate protein structure predictions at the CASP14 conference. We explored network architectures incorporating related ideas and obtained the best performance with a 3-track network in which information at the 1D sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The 3-track network produces structure predictions with accuracies approaching those of DeepMind in CASP14, enables rapid solution of challenging X-ray crystallography and cryo-EM structure modeling problems, and provides insights into the functions of proteins of currently unknown structure. The network also enables rapid generation of accurate models of protein-protein complexes from sequence information alone, short circuiting traditional approaches which require modeling of individual subunits followed by docking. We make the method available to the scientific community to speed biological research.

https://www.biorxiv.org/content/10.1101/2021.06.14.448402v1

---

Very interesting also this part:

We refer to these networks, which also generate per residue accuracy predictions, as RoseTTAFold. The first has the advantage of requiring lower memory (e.g. 8GB RTX2080) GPUs at inference time and producing full sidechain models, but it requires CPU time for the pyRosetta structure modeling step. For proteins over 400 residues, the end-to-end version requires 24G TITAN RTX GPU; we are currently extending this network to generate side chains as well.

:-O

---

Thanks. I hadn't read the full report myself. I hope they can scale it down for us.

But here is a tweet discussing it:

Baker lab's effort at reproducing AlphaFold2 is out on bioRxiv. Pretty impressive performance gains (relative to original trRosetta), if not quite yet at AlphaFold2 level.

https://twitter.com/MoAlQuraishi/status/1404891377419366401

---