AI just designed molecular scissors from scratch ✂️

Most medicines work by holding on. A drug or an antibody binds its target and blocks it, one molecule at a time, so a million copies of a disease protein need roughly a million copies of the drug. Enzymes play a different game: they change their target, let go, and move on to the next one. Last week, the Baker lab at the University of Washington's Institute for Protein Design published two papers in Nature showing that AI-driven design can now build enzymes that cut proteins at a spot you choose. That has been a long-standing open problem, and I find it genuinely exciting.

Why cutting proteins is so hard

The bonds that hold amino acids together are extremely stable: in neutral water a peptide bond can last for hundreds of years. Natural proteases get around that with a "blade" positioned with near-atomic precision, plus a subtle flexibility to grip the right sequence. Getting both right in a protein designed from nothing has been out of reach for a long time.

The teams used deep-learning design tools: a version of RFdiffusion to build the backbones, ProteinMPNN to pick sequences, and AlphaFold 3 to filter the candidates before anything went into the lab.

Two kinds of scissors

Cysteine proteases (led by Hojae Choi) use an activated cysteine as the blade. 13 of 69 designs worked (19%), with rate enhancements of up to 3 × 107 over the uncatalysed reaction. Their folds look nothing like natural proteases, yet six crystal structures matched the design models to within 1.2 Å. And they cut their targets inside living human cells.

Metalloproteases (led by Anqi Chen and Kejia Wu) use a zinc ion instead. Of 135 designs, 36% cut their target peptide at the intended bond. The best one speeds up the reaction more than 100-million-fold, and four mutations push that past 1010. The team then designed proteases that selectively cut disease-related proteins: TDP-43 (linked to ALS and frontotemporal dementia), amyloid-β (Alzheimer's) and serum amyloid A, with rate enhancements of up to 109. They also built "caged" signalling proteins that only switch on when the matching designed protease cuts them free.

Why this is beautiful

What I love here is the shift from AI that predicts protein shapes to AI that designs proteins that do chemistry. David Baker put it as moving to "designing machinery that physically changes matter at the atomic level." Because one enzyme can cut thousands to millions of targets over its lifetime, this could eventually mean lower doses, and a way to reach proteins that offer no pocket for a classic drug.

To keep it honest: this is lab work, not a treatment. The disease targets were tested as peptides or purified proteins (including full-length TDP-43 in a test tube), and the authors note their designs are still less active than the best natural cysteine proteases. But the code is open, the structures are in the Protein Data Bank, and the success rates are high enough that other labs can build on this right away.

Programmable scissors for proteins, designed with AI: that's a sentence I wouldn't have believed a few years ago. 🧬

Sources

Reacties

Een reactie posten

Populaire posts van deze blog

SpaceX just asked to fly Starship hundreds of times a year 🚀

Last week the U.S. Federal Aviation Administration took a quiet but important step: it opened a formal review of SpaceX's request to launch Starship from Starbase in South Texas up to 456 times per year . Today the site is cleared for only 25 flights a year. That jump is the clearest signal yet that Starship is meant to become routine hardware, not a rare test rocket. What the FAA is reviewing On 14 October 2026 the FAA published a Notice of Intent to prepare a Programmatic Environmental Impact Statement (PEIS) for SpaceX's Starbase plans. A PEIS looks at a whole program of operations rather than a single launch, which is exactly what you need if you want a reusable rocket to fly the way airlines fly planes. SpaceX's proposal comes in two phases, according to the Federal Register notice: Phase 1: up to 100 launches and 200 landings per year (100 Ship, 100 Super Heavy), any day or time of week. Phase 2: up to 456 launches and 912 landings per year at Starbase. La...
Meer lezen

AI that can actually "see" a crystal 🔬

Most of the solid stuff around us, from battery materials to chips, is made of crystals: atoms stacked in neat, repeating 3D patterns. Large language models are great at reading papers about those materials, but they have never really looked at the structures themselves. A team at MIT just showed a clever way around that, and the trick is one I keep coming back to: good synthetic data. What they built The model is called CLaC, short for Contrastive Language-Crystal model, and it is described in a paper in Nature Communications by Yang Jeong Park, Mayank Kumaran, Chia-Wei Hsu, Elsa Olivetti and Ju Li. MIT News covered it on 18 October 2026. If you know CLIP, the model that taught image generators to connect pictures and words, you already get the idea. CLaC has two branches: a graph neural network that reads the 3D crystal structure, and a text encoder that reads a description. Contrastive training pulls matching structure-text pairs together in one shared space and pushes mismatche...
Meer lezen

Google's Deep Review wants to be the second pair of eyes on every long document 🔍

AI has made writing fast. Checking has not kept up. That gap is exactly what Google Research is going after with Deep Review, which went public on 14 October. It's an agentic system that reads a long technical document and tries to find what's wrong with it: broken proofs, logic gaps, unsupported claims. For anyone working on legal and document AI, like me, this is a launch worth a close look. From conference pilot to public tool Deep Review is the grown-up version of Google's Paper Assistant Tool (PAT). Over the past year, authors of more than 16,000 manuscripts at STOC 2026, ICML 2026, NeurIPS 2026 and ICLR 2027 opted in for private feedback before the submission deadline, fully separate from official peer review. The need is real: Google counts over 74,000 submissions to ICLR, NeurIPS and ICML this year, up from about 23,000 in 2023. Now anyone can use it. Every user gets two free reviews per week, and you can connect a Gemini API key for higher limits and programmati...
Meer lezen