Profile avatar
alexechu.bsky.social
Deep learning & protein design. Currently designing stuff at Google DeepMind, previously at Stanford with Possu Huang. alexechu.github.io
6 posts 1,425 followers 104 following
Regular Contributor

Bluetorial: Adventures with tenure-Part 1

Many industrially-important reactions (e.g. lipase-catalyzed biofuel production) involve substrates poorly soluble in water. How can we screen enzymes for these reactions? Samuel Thompson invented a new 'triple emulsion' platform for biphasic screening (1/n): onlinelibrary.wiley.com/doi/full/10....

At neurips this week - hmu if you’re around, want to catch up, try out some food, explore, etc!

Latent diffusion ftw!

Super excited to preprint our work on developing a Biomolecular Emulator (BioEmu): Scalable emulation of protein equilibrium ensembles with generative deep learning from @msftresearch.bsky.social ch AI for Science. www.biorxiv.org/content/10.1...

Evolution is non-linear, the next step in the evolutionary lineage of the ESM series of models is a sequence only protein language model that sets a new state of the art in representation learning. Check it out! The 300M and 600M models are open weights, 300M available for commercial usage.

AI Proteins is excited to partner with @bms-news.bsky.social to develop novel miniprotein-based therapeutics. This collaboration will leverage our cutting-edge AI-driven platform to design and optimize miniproteins with the potential to transform patient outcomes www.businesswire.com/news/home/20...

“On a log-log plot, my grandmother fits on a straight line.” -Physicist Fritz Houtermans There's a lot of truth to this. log-log plots are often abused and can be very misleading 1/5

AF3 BEST METHOD followed by cluspro but also some conversion errors

go.bsky.app/BoXpqQp Computational protein design starter pack. Let me know if I missed you!

Asking the following earnestly: what is the strongest case for GANs standing the "test of time"? Are they important 10 years later in modern ML research? How have they influenced the way we think about generative models today?

A great new essay on AI for Science from our colleagues here: deepmind.google/public-polic...

Amazing blog post on flow matching, stunning visuals! It also makes the connection with normalising flows crystal clear. Incredible effort!

A weekend project from a while back -- this little package (with no dependencies) allows you to interact with pymol remotely. I use it a lot for my protein design workflows together with @biotite.bsky.social. Just `pip install pymol-remote`

The complete guide for transfer learning with the protein language model ESM-2. (In brief: Use ESM-2 650M and calculate mean embeddings across sites.) www.biorxiv.org/content/10.1...

Experimental kcat and KM measurements for hundreds of naturally variants of adenylate kinase. - thermophilic enzymes are not slower than mesophilic! - general kcat/KM predictors are bad and easily beaten by models trained on this specific dataset! www.biorxiv.org/content/10.1...

Our team at deepmind is hiring! boards.greenhouse.io/deepmind/job...