I'm very late to sharing this news: Deep Learning for RNA design is published in Science and chosen as the cover article! Now over 3 years ago, gRNAde started as a side-project in my CS PhD because I wanted to do something real beyond benchmarks. Its been an amazing journey with @RDasLab @eternagame @HHMIJanelia @pl219_Cambridge, from new ML modelling, to an actual computational design system, to it working experimentally, to Science!

🚀 We’re hiring at the newly established Singapore Embodied AI Centre @eaic_sg ! 🇸🇬🤖 Open roles: AI/Robotics Research Scientists, Tech Leads, Software Engineers, and Full-Stack Roboticists. Excited to also join as part of the inaugural cohort of Affiliated PIs! Come build the future of robotics with us!

OpenKnot round 9 results will be revealed Thursday, November 5, 8:30am PT during an Eterna town hall meeting, followed by @AF_Panescu describing the Struct2SeQ-OA method for inverse RNA design she used for her designs. Details at eternagame.org/news/16569676

🚀 Excited to share BindClaw! 🦀 Starting from a prompt or research paper, BindClaw connects biological mechanisms to protein design: identify targets and binding sites → build mechanism-specific computational assays → generate and assess binders → refine designs using the results. We use agents to evaluate whether designs fit the intended biological objective, including competition with a native partner and preference for a particular molecular state. 🧬 In a human-guided calmodulin campaign, 15 of 18 designs showed binding to calcium-loaded CaM, including subnanomolar binders supported by both BLI and SPR. This work is an incredible team effort together with @DidiKieran , @saint__ai, @ Kyle Gion and @sacdallago and many others! 📄paper/project: research.nvidia.com/labs/dbr… 💻code: will be released in a few weeks!

In new work, we report a surprising finding: prompting Evo 2 with a handful of sequences of shared function can steer the model to generate new sequences with similar function, a form of generative in-context learning (ICL). This is an emergent phenomenon that we did not anticipate during training. In a first study, led by Alex Hao, we show that ICL enables few-shot design of mammalian regulatory elements, including novel enhancers with cell-type specificity for a neuronal subtype in the mouse brain. In a second study, led by @garykbrixi and Alex Hao, we extend these results to the computational design of natural proteins, de novo antibodies, protein-RNA systems, and complete bacteriophage genomes. As part of this work, we also release several useful new artifacts, including a highly efficient new inference stack. 1/n

Stellar lineup at the AITHYRA Next generation AI for Biomedical Science workshop -- next Monday 2 Nov at the IMBA lecture hall in Vienna Biocenter. No registration is required.

my hypothesis for why AI-generated slides for technical presentations are so awful: people who don't have a lot of experience presenting think that to make a talk you just put all of the stuff you want to talk about on the slides, then you go through each slide and say the stuff the slide says. how you're actually supposed to prepare a presentation: you decide what to say, write it out as a rough script, then put maybe 3% of the text on the slide, plus a bunch of visual aides (e.g. plots that only include data you're planning to discuss). AI-generated presentations are mostly the former. they have a ton of stuff the presenter never mentions, and the stuff that is most useful for the presenter is absent or not highlighted. they include enough content for a presentation 10 times longer than the actual presentation is supposed to be. maybe try doing it the other way: first write the script for what you actually want to say, and make sure you can say all of it in the time you have. then prompt the model to create the slides based on the script. ask it to use as little text a possible, and a minimalist visual style rather than the claude house style. you can thank me later (I accept both venmo and zelle).

it almost seems that there are two entirely separate frontiers of comp protein design right now: a) an entire new class of chemistry is now available for computational enzyme design b) we made ML-designed VHHs bind 5% better 🙂

Introducing Recurse, a small personal project I built for my own use: a research-focused, minimal agent harness built on Pi Durable. I built it autonomously by managing a team of Opus and Sol agents, in an effort to make intelligence my own. Every part is open source, so you can make it yours too :)

Wow! And best wishes to Bowen and Divergent team! 🚀

The best way to identify and work through the bottlenecks of drug development is to get deeply involved in real drug programs, across many different therapeutic areas and stages of development. Over the past several months, our scientists have engaged in deep research partnerships with six organizations developing new medicines: The Column Group, Generate Biomedicines, BigHat Biosciences, Divergence Labs, Enveda, and Bristol Myers Squibb. We’ve also begun our own drug discovery efforts through the acquisition of Mithril Medicine, a single-person biotech company co-founded by @atlasventure. Today we’re sharing early progress in accelerating drug discovery from these research collaborations. With Claude, we are streamlining operations and working to derisk novel targets with TCG, redesigning lab protocols with BigHat, designing CAR-T cell therapies with Generate, and multi-specific therapeutic antibodies with Divergence, drafting regulatory filings with Enveda, and supporting portfolio-level decision making with BMS. Our early collaborations span pre-clinical and clinical work on potential treatments for cancer, allergies, rare genetic diseases including forms of childhood muscular dystrophy, and more. My team has been working with @Prof_Oak_ , founder of Mithril Medicine, for the past year. Jason Rhodes, General Partner @atlasventure and Chairman of the Mithril Board said this about Mithril: “We prioritized LAMA2-Related Congenital Muscular Dystrophy (LAMA2-RD) as the lead indication, in part, because of the enormous unmet need. These patients present with the disease at birth, spend their lives in wheelchairs and on ventilators, and rarely live past their mid-twenties; yet no therapy is available, and few are in active development. We believed Mithril would be the company to change that.” Anthropic acquired Mithril and we are pushing the frontier of how a single skilled operator, with Claude, can run an entire pipeline of drugs. We’re working with these teams to show how AI can assist drug developers at many stages of the process already, and to prioritize improvements to our models based on the benchmarks we’re building with collaborators. Our blog includes links to technical reports written with our partners: anthropic.com/research/early…

Very excited to report that Mint has published an in-depth profile of our work @PopVaxIndia as part of @_jesthetic's wonderful new biotech series "Lab Rats". Read it here: livemint.com/companies/start… Looking forward to future installments! (1/2)

Really interesting podcast - highly recommend (and also really approachable, not intense/overly academic!)

How Demis Hassabis talks about doing your best… Its really something extraordinary and resonates with me

📢 RFD4-Proteina talks 📢 Great seeing @kpowgerade introduce RFD4-Proteina at @nvidia GTC Berlin, and fun presenting it with @nscorley at the GTC Biopharma summit! Next up: our own little Eurotour - with guest appearances from David himself in the UK. Come by if you’re around! 🧵 1/3

Please reach out if you're interested in working on ambitious molecule design problems together!

Today we are releasing UMA Nitro! ⚡ Nitro brings unprecedented inference speeds to the latest UMA models – reaching 6 ns/day on 1k atoms on a single H200 GPU, and up to a blazing 33 ns/day raw inference speed on ~150 atoms! Try it now - github.com/facebookresearch/…!

Delighted to share a landmark preprint from our lab, using genome language models and AI agents for biological discovery and genome mining! The stretches of DNA between protein-coding genes are often called the genome’s “dark matter,” but they are far from empty. They encode RNA molecules that sense metabolites, regulate genes, and control some of molecular biology's most important tools (such as programmable genome editors e.g. CRISPR/bridge) Much of this biology is unexplored. One reason is that RNAs can retain similar structures and functions even as their DNA sequences change dramatically, making their relationships difficult to recognize with conventional computational tools

Today, we’re announcing an open call for guest investigators in life sciences. Over 8–12 weeks, guest investigators will work directly with my team and other Anthropic researchers to produce analyses and tools for their work. We’ll run our models and harnesses, at no cost to them. We’ll work together to open-source the resulting data, artifacts, and code, and write preprints about what we find. We want humans to remain at the center of scientific discovery, deciding what questions to pursue, and how to pursue them. Those decisions are not ours to make alone. This open call is part of our ongoing efforts at Anthropic to support researchers and accelerate scientific discovery. I’m looking for principal investigators and their research groups at academic institutions, nonprofits, or government labs, working in causal human biology, structural and computational molecular biology, medicinal chemistry, drug discovery, and clinical science. Applications are open for about 14 days: anthropic.com/news/open-call…

I have a lot of people to be thankful for in life, but @SimMat20 is definitely very high on the list, for being an amazing coworker, friend and introducing me to equally amazing people like @nscorley. Cannot imagine not working with this guy anymore!