anybody in AI x bio who wants to know what industrial drug discovery is actually about should read this briefing by Treeline Bio (announced $200m additional capital yesterday, have raised $1.1bn total)
led by serious drug hunters Josh Bilenker and Jeff Engelman
link in…
Curious what folks who work on iPSCs think of the @OpenAI x @RetroBio_ [*] collab to design efficient Yamanaka factors using a fine-tuned GPT-4o.
I wonder to what degree the designs integrate all the previously published work on sequence properties of improved Sox2/Klf proteins
Boltzina: Efficient and Accurate Virtual Screening via Docking-Guided Binding Prediction with Boltz-2
1. Boltzina is a novel framework that combines the high accuracy of Boltz-2 with the speed of AutoDock Vina for virtual screening in drug discovery. It achieves this by directly…
🎙️ Introducing TechBio Talks – a new podcast hosted by @RecursionChris and Najat Khan that offers candid conversations with the scientists, founders, and investors on the frontlines of the AI transformation of drug discovery.
💥 TechBio Talks kicks off after Labor Day –…
Have protein–ligand cofolding methods moved beyond memorization?
It is well recognized in AI research that many models excel at memorizing patterns from training data, but struggle to generalize to truly novel cases. This limitation is especially consequential in drug discovery,…
🚨 new paper alert 🚨
We present a method for designing protein nanoparticles with two unique faces that can bring together distinct biological targets at controlled distances!
This project was led by Dr. Sanela Rankovic is out now in Nature Materials: tinyurl.com/yj8t476d
Multi-state Protein Design with DynamicMPNN
1. DynamicMPNN is a groundbreaking inverse folding model that generates protein sequences compatible with multiple conformations. It outperforms ProteinMPNN by up to 13% on structure-normalized RMSD, demonstrating significant…
Generative Design of High-Affinity Peptides Using BindCraft
1. A groundbreaking study leverages BindCraft, an AlphaFold-based platform, to design high-affinity peptide ligands directly from protein structures. This approach marks a significant step forward in de novo peptide…
Thrilled to share our new paper in @ScienceMagazine! 💥
We developed a computational platform to design tiny 'minibinders' from scratch that can guide T cells to kill cancer cells. This is proof-of-concept for targeting intracellular cancer antigens with a new modality! 🎯
🤯Three (!!!) new papers today in @ScienceMagazine on the application of generative AI for the de novo design of peptide/HLA binding molecules! Completely unique 3D structure and binding mode compared with natural TCRs and TCR mimics! Links to papers 👇
Cold tumors–cancers with little to no immune cell infiltration–represent one of oncology’s toughest challenges. These tumors evade the immune system, leaving patients with limited treatment hope.
What if we could turn cold tumors “hot”? Published in Springer Nature…
⚛️Quantum meets protein design🧬
QOBRA = a variational quantum auto‑encoder w/ 27 trainable angles + 9 qubits
... and still designs Ca²⁺/Mg²⁺/Zn²⁺ proteins that Chai‑1 + MD say will fold & bind
last couple days:
La-Proteina
ATOMICA
BioScore
Metalorian
AMix-1
SO(3)-DiT
EZPred
AlphaFlex
Ibex
yet we still don't have a standard eval framework for these generative models that allows ppl to identify which is optimal for their task. are ppl just picking models off vibes?
Boltz-2 on Novel Ligands for a Known Target
At KlyneAI, we believe model performance should be evaluated under the toughest conditions - not just retrospective benchmarks.
In this case study, we challenged Boltz-2 with a set of novel ligands against a well-characterized PDB…
We just open sourced @ProjectBiomni! Checkout our github repo github.com/snap-stanford/… and we are very excited to build our E2 environment together with the community!
We just open sourced @ProjectBiomni! Checkout our github repo github.com/snap-stanford/… and we are very excited to build our E2 environment together with the community!
We discovered:
🧬 Complete E–M protein models
🫒 3 lipid-binding pockets
☠️ Mutating one tryptophan (W467) destroyed the virus
That residue is conserved across flaviviruses
→ A promising pan-flaviviral drug target (3/5)
2K Followers 284 FollowingImproving life science with programmable experiments. 🧬 🤖 Home of the Protein Engineering Tournament & Open Datasets Initiative
233 Followers 371 FollowingResearch Scientist at ByteDance Research. Yes we do some #cryoem (algorithm development). Ex @cianfrocco_lab and @JulieBiteenLab. Opinions are my own.
194 Followers 507 Followingmolecular biophysics phd @ucsandiego
working on bio+ml, generally interested in proteins
documenting my research journey / posting thoughts on papers
893 Followers 6K FollowingHyperdidactist.
Philologist; Obstinated-Criticist.
Support Ecology-Economy of Minds
PhD ENGINEER {semantic,graphic}memory
IN SCIENCE We Trust. #TheoryWave
1K Followers 2K FollowingPostdoc with Debora Marks. Previously MIT EECS PhD. Starting FutureHouse Fellowship (in SF for September). Come talk to me about modeling viral/immune proteins!
632 Followers 562 FollowingProfessor at the School of Biomedical Sciences, University of Hong Kong. Uses genomics data to understand DNA repair and cancer.
858 Followers 449 FollowingProtein biochemist working on structural aspects of gene regulation. For (more) fun, I run, rogaine, bike tour and play tenor sax. @Sydney_Uni @SydneySOLES
2K Followers 284 FollowingImproving life science with programmable experiments. 🧬 🤖 Home of the Protein Engineering Tournament & Open Datasets Initiative
6K Followers 980 FollowingAssistant Professor @MITChemE & @MITEECS | ChE, Chem, AI/ML, etc. for accelerating molecular synthesis and discovery | https://t.co/cAcVl8CjqR | he/him
108K Followers 1 FollowingClaude is an AI assistant built by @anthropicai to be safe, accurate, and secure. Talk to Claude on https://t.co/ZhTwG8dz3D or download the app.
32K Followers 2K FollowingMy posts are strictly personal, not reflective of positions or policies of Insilico Medicine, which disclaims any responsibility for my individual statements.
11K Followers 145 FollowingThe official handle for #NVIDIAHealthcare. Helping the scientific and developer community advance research, diagnostics, and patient care with #AI.
8K Followers 477 FollowingInsilico Medicine is a clinical-stage AI-driven biotech company dedicated to extending human productive longevity by transforming drug discovery with AI.
16K Followers 580 FollowingScientist (PI). Doing fundamental research in #Materials and #Electrochemistry. Ex-PSI. Ex-Stanford. Run a Podcast (on Youtube).
2K Followers 460 FollowingMolecular Parasitologist. Teacher. Fascinated by how parasites interact with the immune system. Rational vaccine designer. Country boy and nature lover. 🏳️🌈
396 Followers 798 FollowingGroup Leader at the Oxford Biochemistry Department interested in molecular mechanisms underlying viral evolution and host changes.
4K Followers 461 FollowingFollow for AI in Digital Biology and Drug Discovery @NVIDIA, ex Insilico Medicine, ex Yale, PhD UMaryland, views are mine, DM for collabs
841 Followers 42 FollowingStudent-run account for the Green group (MIB).
Enzyme design. Directed evolution. Genetic code expansion. Industrial biocatalysis.
885 Followers 687 FollowingSenior Lecturer in Biotechnology, University of Edinburgh. We develop methods and tools to make protein design and engineering more reliable and accessible.
2K Followers 363 FollowingNovel biologics designed from natural proteins using Rosetta, AI, and large scale lab screening. Pipeline: Severe autoimmune disease, COVID, metabolic.