RT @james_y_zou: What is it like for scientists to work with AI co-scientists? . Excellent @Nature feature discussing researchers' experien….

RT @james_y_zou: >700K people die each year due to S. aureus infection. Today we show that our AI designed new molecule, synthecin, stops….

RT @peholderrieth: Our MIT class “6.S184: Introduction to Flow Matching and Diffusion Models” is now available on YouTube!. We teach state-….

Very cool work from @jaanakprashar, @yusufroohani, and others at the @arcinstitute!.

RT @CircAHA: ADMET-AI accurately predicts drug-induced cardiotoxicity and allows interpretation via most important ADMET properties: implic….

RT @james_y_zou: Our new paper in @CircAHA shows how #ADMET-AI predicts cardiotoxicity of drugs and generates interpretable insights https….

RT @Joseph_C_Wu: Our latest review, "Next-Gen Therapeutics: Pioneering #DrugDiscovery with #iPSCs, #Genomics, #AI, and Clinical Trials in a….

Thank you @kevinroose for highlighting SyntheMol! SyntheMol is our generative AI model for drug design that ensures easy synthesis of generated molecules, allowing us to rapidly identify six AI-designed molecules with promising antibiotic activity.

RT @james_y_zou: 📢 Excited that #unitox is selected as a #NeurIPS2024 spotlight!💡. We created #LLM agent to analyze >100K pages of FDA docs….

The Virtual Lab architecture can easily be adapted to other scientific domains as well. The code is freely available at Thanks to Wesley Wu, Nash Bulaong, John Pak, and @james_y_zou for the great collaboration!.

Our collaborators at the @czbiohub experimentally validated 92 nanobodies designed by the Virtual Lab. Two of these designs—one mutant Nb21 nanobody and one mutant Ty1 nanobody—gained binding to recent variants such as JN.1 and KP.3.

As a real-world case study, we used the Virtual Lab to design nanobodies to bind to recent variants of SARS-CoV-2. Through a series of team and individual meetings, the Virtual Lab built a computational nanobody design pipeline utilizing ESM, AlphaFold-Multimer, and Rosetta.

The Virtual Lab consists of a human researcher and several LLM agents, including a PI, a scientific critic, and several science agents. The human and LLM agents discuss science via team meetings with all the agents and solve tasks via individual meetings with specific agents.

I'm excited to share our recent work on the Virtual Lab, a collaboration between human researchers and LLM agents for science research. We apply the Virtual Lab to design nanobody binders for recent variants of SARS-CoV-2 with experimental validation.

Very cool @NewYorker article by @DhruvKhullar on AI for antibiotic discovery. Thank you for covering some of our work with @ItsJonStokes and @james_y_zou! References below. Chemprop: SyntheMol:

RT @james_y_zou: Very excited to share our primer on LLM for biology @naturemethods! We discuss how to use:.⚡️human & protein LMs (eg #ESM)….

I'm excited to share our @naturemethods primer on language models for biological research. @ElanaPearl, @james_y_zou, and I explain how to use language models, both natural language models and biological language models, to accelerate biological research.

RT @james_y_zou: 📢#AI predicting molecule properties to guide drug discovery. Our updated ADMET-AI paper is now published:.📰 Paper: https:….

ADMET-AI is now published in Bioinformatics! ADMET-AI provides free, fast, & accurate prediction of absorption, distribution, metabolism, excretion, & toxicity properties of molecules at Blog: Paper:

RT @ElanaPearl: The Illustrated AlphaFold. Do you want to know how AlphaFold3 works? It has one of the most intimid….