🧬📈 How to optimize proteins using ML models - Part 1 For this blog post, we surveyed the state of the art ML methods for optimizing proteins such as antibodies and enzymes. We focused on adaptive methods, suitable for cost constrained campaigns. Check it out here:

Very useful evaluation of machine learning strategies for #proteinengineering #directedevolution by Francesca-Zhoufan Li @ ICLR'25 Yisong Yue Jason Yang Kadina Johnston and coworkers. #MLDE #machinelearning #AIforproteins

The #AlphaFold 3 model code and weights are now available for academic use. We Google DeepMind are excited to see how the research community continues to use AlphaFold to address open questions in biology and new lines of research. github.com/google-deepmin…

Come join our London wet-lab team! Deadline to apply is next Wednesday boards.greenhouse.io/deepmind/jobs/…

Thrilled to share that EVOLVEpro is now published Science Magazine Since our preprint, we now demonstrate low-N eningeering of both antibody and enzymes. We hope this model will broadly useful to the protein engineering field. science.org/doi/10.1126/sc…

A foundation model for enzyme retrieval and function prediction is released. EnzymeCAGE is trained on ~1M enzyme-reaction pairs via CLIP, and can be used to annotate unseen enzymes or orphan reactions. We open-source it before releasing the manuscript: github.com/GENTEL-lab/Enz…

The White House web site with all the Biden PCAST products has been wiped clean. But, thanks to Eric Horvitz, you can still find all our reports at erichorvitz.com/PCAST_reports_….

Latent Labs comes out of stealth today with $50M funding. Our goal? To push the frontiers of generative biology, giving partners instant access to tools capable of accelerating drug design. Every biotech or pharma company searching for the best therapeutic molecules understands

Structural biology is in an era of dynamics & assemblies but turning raw experimental data into atomic models at scale remains challenging. Minhuan Li and I present ROCKET🚀: an AlphaFold augmentation that integrates crystallographic and cryoEM/ET data with room for more! 1/14.

Our team has a Student Researcher position available for this year! probably in something related to protein modeling & design. dates flexible, up to 20 weeks, in sunny mountain view - apply here if interested or reach out with any questions: google.com/about/careers/…

We present InverseBench, a framework for benchmarking plug-and-play diffusion approaches for inverse problems in physical sciences. (#ICLR2025 Spotlight) PnP diffusion approaches are attractive for scientific inverse problems because they offer flexibility in incorporating

It was a pleasure to contribute to this report by National Academies on AI and biosecurity. AI has enormous potential to improve biological research and human health, especially if done so safely and responsibly!

Protein dynamics was the first research to enchant me >10yrs ago, but I left in PhD bc I couldnt find big expt data to evaluate models. Today w Gina El Nesr, I'm thrilled to share the big dynamics data I've been dreaming of, and the mdl we trained: Dyna-1. rb.gy/de5axp

🧬 Meet Lyra, a new paradigm for accessible, powerful modeling of biological sequences. Lyra is a lightweight SSM achieving SOTA performance across DNA, RNA, and protein tasks—yet up to 120,000x smaller than foundation models (ESM, Evo). Bonus: you can train it on your Mac. read

Happy Sunday to all! This morning, we are excited to share Chase’s work developing a simple, scalable method to assemble 100s-1000s of custom genes from oligo pools using standard lab tools! (Small 🧵 below)

Check out this work led by the incredible Ada Fang ATOMICA constructs representations of molecular interactions at the atomic scale, capturing universal physiochemical principles. Read about the exciting insights from applying ATOMICA below 👇

In a medical milestone, a customized base editor was developed, characterized in human and mouse cells, tested in mice, studied for safety in non-human primates, cleared by U.S. FDA for clinical trial use, manufactured as a complex with an LNP, and dosed into a baby with a severe,

I'm excited to share our new preprint "Steering Generative Models with Experimental Data for Protein Fitness Optimization" (🧵1/5)! Paper: arxiv.org/abs/2505.15093 Code: github.com/jsunn-y/SGPO

Guide your favorite protein generative model with experimental data? Meet ProteinGuide - a method to condition pre-trained models on properties without retraining. We validated it both in silico by guiding ProteinMPNN and ESM3 on 3 tasks and in vitro by engineering base editors.

🚀 Excited to release BoltzDesign1! ✨ Now with LogMD-based trajectory visualization. 🔗 Demo: rcsb.ai/ff9c2b1ee8 Feedback & collabs welcome! 🙌 🔗: GitHub: github.com/yehlincho/Bolt… 🔗: Colab: colab.research.google.com/github/yehlinc… Sergey Ovchinnikov Martin Pacesa