Sebastian Swanson

Kseniia Petrova, a Russian scientist at Harvard Medical School, was detained at Logan Airport after returning from France and sent to an ICE detention in Louisiana, The Insider reports.

A big critic of Putin and the war in Ukraine, she may now be deported to Russia.

Have protein-ligand co-folding methods moved beyond memorisation? Deep learning has driven major breakthroughs in protein structure prediction, however the next critical advance is accurately predicting how proteins interact with other molecules, especially small mo...

Excited to share our latest preprint evaluating AlphaFold3, Boltz-1, Chai-1 and Protenix for predicting protein-ligand interactions, featuring our newly introduced benchmark dataset 🌹Runs N’ Poses🌹!

www.biorxiv.org/content/10.1...

🧵👇 (1/n)

Jointly Embedding Protein Structures and Sequences through Residue Level Alignment Residue Level Alignment integrates protein sequence and structure information in a self-supervised model, improving speed and precision in predicting protein binding and structural stability.

✨new work from our lab!✨

RLA is a contrastive-learning approach that assesses sequence-structure compatibility by aligning sequence and structure machine learning representations! RLA can successfully filter protein binder designs. 1/3 🧶

journals.aps.org/prxlife/abst...

schedule introduction and membership Boston Protein Design and Modeling Club (BPDMC) is a community of computational protein engineers and modelers from both academia and industry. While we are based in Boston...

There's just one meeting left in 2024, and we have a spectacular seminar by Mireia Solà Colom lined up for you! Join us at 7pm in Room 181, Building 68, MIT on Wednesday, December 11th.

"De novo design of miniprotein-based natural killer cell engagers"

See you there & Happy Thanksgiving!
bpdmc.org

Design of facilitated dissociation enables control over cytokine signaling duration
www.biorxiv.org/content/10.1...

Come join the protein engineering department at Novo Nordisk! This position is based out of Lexington, Mass. Feel free to reach out to me if you're interested/have questions about the role.
www.novonordisk.com/careers/find...

That's actually super intriguing to me and a little hard to rationalize. Are most of the generated proteins at length ~1,000 still a single domain? I assume so if you're using TM-score, but that could be a confounding variable

How many structures did you generate? It looks like once you get into the > 600 residue regime, Chroma->partialRFdiff->MPNN outperforms all other methods.