Reposted by Amir MotmaenAlex Rubinsteyn @alexr.bsky.social · 15/07/20265.6 is good and Fable refuses to do bio Might pause my Claude subscription for a while 391
Reposted by Amir MotmaenRob Patro @robp.bsky.social · 15/07/2026It's not just bio in my experience; Fable is broadly useless for a large swath of reserach including CS theory, certain math, all bio (obviously), straight up SWE that touches any bio /security/other sensitive things, chemistry, etc. Fable is worse than 5.6 and even Opus, as it's basically absent. 031
Reposted by Amir MotmaenAlex Rubinsteyn @alexr.bsky.social · 21/11/2025Hey @jamieheather.bsky.social what do you think of this? www.biorxiv.org/content/10.1...biorxiv.orgTargeting peptide-MHC complexes with designed T cell receptors and antibodiesClass I major histocompatibility complexes (MHCs), expressed on the surface of all nucleated cells, present peptides derived from intracellular proteins for surveillance by T cells. The precise recogn... 184
Reposted by Amir MotmaenNate Corley @ncorley.bsky.social · 15/08/2025(1/7) Training biomolecular foundation models shouldn't be so hard. And open-source structure prediction is important. So today we're releasing two software packages: AtomWorks and RosettaFold3 (RF3) [www.biorxiv.org/content/10.1101/202…...)biorxiv.orgAccelerating Biomolecular Modeling with AtomWorks and RF3Deep learning methods trained on protein structure databases have revolutionized biomolecular structure prediction, but developing and training new models remains a considerable challenge. To facilita... 26728
Reposted by Amir MotmaenFlorian Praetorius @flopraetorius.bsky.social · 28/11/2023The Praetorius lab for Biomolecular Design at the Institute of Science and Technology Austria (ISTA) is looking for grad students in 2024. If you are interested in protein design at a great institute near Vienna reach out to me! www.dropbox.com/scl/fi/6iny2...dropbox.com 1118
Reposted by Amir MotmaenSergey Ovchinnikov @sokrypton.org · 16/11/2023Late night experiment with @chrisfrank662.bsky.social :D The backbone designability test. Start w/ backbone (diffusion, hallucination), put a sequence on it (proteinmpnn, esm-if), then the final test is to make sure it folds back to the original structure (alphafold, omegafold, esmfold) (1/3) 21912