Sabitlenmiş Tweet
Rebuilt the latest PyRosetta4(v2026-02-06), finally fixed the GIL(Global Interpreter Lock)
github.com/ullahsamee/PyR…
English
sk
355 posts
@ai4s_protenix here's the secret sauce
Mohammed AlQuraishi@MoAlQuraishi
We're also releasing a massive self-distillation set, a key ingredient in training AF3, comprising millions of diverse MSAs and structures. We estimate its cost to near $20M, representing perhaps the largest compute investment by an academic effort for a biological dataset. 6/9
English
@AsimovPress @btnaughton AdaptvybBio competition i think used ipsaemin not ipsae(dunbracks) right?
English
A few years ago, designing an antibody on the computer was extremely difficult.
Today, there are several open-source tools which allow anyone to design antibodies from home.
Out today: A step-by-step guide to antibody design. By @btnaughton.
English
Zerocam II for Android is all about pure camera power and the absence of complexity.
Available very soon.
English
English
Yo @zerocam_app just released 2.0 for Android, excited to give it a try on my @nothing Phone (3) 🔥
English
@compchemm Cool! But I see no wheel yet in the Releases page?
github.com/ullahsamee/PyR…
English
Rebuilt the latest PyRosetta4(v2026-02-06), finally fixed the GIL(Global Interpreter Lock)
github.com/ullahsamee/PyR…
English
Comes with latest beta_jan25, an updated energy function, better than beta_nov16. It improves performance on protein-protein interface prediction and reduces atomic clashing.
biorxiv.org/content/10.648…
English
@compchemm Actually, the loop on the Boltz prediction seem to better modeled - AFAIK, it it flexible and this flexibility is captured better.
English
Run Protenix-v1 with Blackwell NVIDIA RTX 5090 Latest PyTorch 2.10, CUDA 13.0 github.com/bytedance/Prot… Boltz2(1st img) versus Protenix-v1
English
Today we share a technical report demonstrating how our drug design engine achieves a step-change in accuracy for predicting biomolecular structures, more than doubling the performance of AlphaFold 3 on key benchmarks and unlocking rational drug design even for examples it has never seen before.
Head to the comments to read our blog.
English
@GregPreibisch Not shown here but when aligned on crystal (Boltz2 0.4Å), (Protenix-v1 0.3Å). Top3 lowest rsmd from Boltz2 has that yellow very low plddt loops. Run Boltz2 (RECYCLING_STEPS=10, DIFFUSION_SAMPLES=25, SAMPLING_STEPS=200, MAX_MSA_SEQS=8192).
English
🚀 Olmo 3.1 is here — earlier than expected!
32B Think: 3 extra weeks of RL training = steady gains and significant improvements.
32B Instruct: Our 7B recipe scaled to 32B, tuned for short chat + function calling.
Olmo 3 keeps leveling up! Details in the latest version of the paper. As always, kudos to the whole team!
New Olmo 3.1 artifacts: huggingface.co/collections/al…
Paper (arxiv soon): allenai.org/papers/olmo3
Demo: playground.allenai.org
Ai2@allen_ai
Olmo 3.1 is here. We extended our strongest RL run and scaled our instruct recipe to 32B—releasing Olmo 3.1 Think 32B & Olmo 3.1 Instruct 32B, our most capable models yet. 🧵
English
Protein Design Labs updated
- 235 groups working on so many interesting protein design projects from 25 Regions worldwide.
- Open Positions Tab now includes > 100 curated #PhD , #postdoc , Masters and internship #OpenPositions link: ullahsamee.github.io/proteindesigng…
GIF
English
Protein Design Labs worldwide - updated
ullahsamee.github.io/proteindesigng…
Please add any groups missing from the list in the issues section. Thank you
#structuralbiology #alphafold #compchem
English