We’ve released an improved version of SynFlowNet! SynFlowNet can now handle large action spaces (up to 220k building blocks) and we experiment with training different GFlowNet backward policies. (1/3) arXiv: arxiv.org/abs/2405.01155 code: github.com/mirunacrt/synf…

Check out our results and discussion in our recently published preprint! (3/3) Thanks Charlie Harris Ilia Igashov Arne Schneuing Marwin Segler Bruno Correia Julien Roy Emmanuel Bengio Pietro Lio'

Researchers led by Ludwig Oncology UNIL CHUV Melita Irving & Greta Maria Paola Giordano Attianese have developed CAR-T cells for cancer #immunotherapy that can be switched on to varying degrees of intensity & then switched off on demand with existing drugs. bit.ly/3YweSL6

Ludwig Cancer Oncology UNIL CHUV I am ever so grateful to my amazing group for their hard work and determination towards our goal of improving the safety and efficacy of CAR T cell therapy of solid tumours. Thanks also to our collaborators Bruno Correia lab and support from Swiss National Science Foundation and Ludwig Cancer!!

📢 Join us next Friday for a talk with Hannes Stärk (MIT) and Bowen Jing (MIT) on MDGEN: Generative modeling of molecular dynamics trajectories Michael Bronstein and Bruno Correia 📅Fri, Nov 15 - 5p CET/11a ET 📝genaiindrugdiscovery.com

Looking forward to the results! There are some seqs in there that I would never order (like `VALLLLLLL`, `IVVMVMVMVM`, `SPAAAAGAA`, `PPPPPPPPPV`). PDBs on the leaderboard (in addition to the seq would be helpful). Good luck to everybody!

🥳 Design proteins from scratch, with a good helix/sheet balance, quickly and robustly with fancy geometric algebra 🤓 that catches how proteins look like! Stellar team work by Leif, Simon, Seva Seva Viliuga and Nico, wonderful HITS Lab project with Jan Stühmer.

🚨Arne Schneuing, Charlie Harris, Yuanqi Du, Michael Bronstein, Bruno Correia and colleagues evaluate how diffusion models can be used to address structure-based drug design problems. nature.com/articles/s4358…

Here is our large-scale protein design story with Bruno Correia and Novo Nordisk! Using geometric deep learning, we evaluated the entire human surfaceome, identified 4,500 sites, and targeted them with a library of 0.5M seeds, equal to performing 3B docking! biorxiv.org/content/10.110…

Mapping Targetable Sites on the Human Surfaceome for the Design of Novel Binders 1. A groundbreaking study maps the human surfaceome, identifying 4,500 targetable sites across 2,886 cell-surface proteins. This resource unlocks new therapeutic opportunities for precision

🔬 New geometric machine learning method accelerates precision drug development! Michael Bronstein, #AITHYRA Director @OeAW and DeepMind Prof at University of Oxford & Bruno Correia's team at EPFL published their study in Nature: go.nature.com/4haJ8S0 ➡️ Read more: bit.ly/3E5DeU0

Our paper on computational design of chemically induced protein interactions is out in nature. Big thanks to all co-authors, especially Anthony Marchand, Stephen Buckley and Bruno Correia nature.com/articles/s4158…

Talking about simulations/structure prediction at UCL. And the quest to share the little I know and get more students into ML for Bio continues. Thank you CS Society team for the invitation!

So glad to be part of this great work! #Nature #EPFL nature.com/articles/s4158…

We’re excited to welcome Nobel Laureate Ardem Patapoutian to EPFL ! Join us for a seminar on Piezo channels and mechanosensory biology. 🗓️ Jan 27, 14:15 📍SV 1717, EPFL ℹ️: memento.epfl.ch/event/ardem-pa… #Mechanotransduction #EPFL #LifeSciences Mackenzie Mathis, PhD

Apart from a twisted ankle while trying to win the sledge 🛷 race, no casualties to report from this year‘s lab retreat in the snow!

We are recruiting🤗! please share (and apply)

New in nature from Arne Schneuing, Stephen Buckley, Bruno Correia, and team: “Targeting protein–ligand neosurfaces with a generalizable deep learning tool.” Congrats on your paper and thanks for choosing Twist DNA to support this work 🧬 go.nature.com/4hjYYd7

🎉👏 Huge congratulations to Mubarak Idris for winning the 2025 Best Master Thesis Award #SNEChemBio! 🏆🎓 His work on “Computational Design of Functional Protein Switches with Therapeutic Potential” sets a new bar for innovation! Faculty of Science | UNIGE EPFL

Computationally designed stem-epitope mimetics elicit broadly reactive antibodies 1. The study introduces a novel approach to vaccine design, focusing on computationally engineered stem-epitope mimetics that aim to elicit broadly neutralizing antibodies (bnAbs) against the