Benefits of AI-Designed Drugs

AI just designed a clinically effective antibiotic that works against MRSA. Most generative models in drug discovery propose molecules that can’t be synthesized or validated. That’s changing. 𝗦𝘆𝗻𝘁𝗵𝗲𝗠𝗼𝗹-𝗥𝗟 𝗶𝘀 𝗮 𝗿𝗲𝗶𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝗳𝗼𝗿 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝗻𝗼𝘃𝗲𝗹, 𝘀𝘆𝗻𝘁𝗵𝗲𝘁𝗶𝗰𝗮𝗹𝗹𝘆 𝘁𝗿𝗮𝗰𝘁𝗮𝗯𝗹𝗲 𝗮𝗻𝘁𝗶𝗯𝗶𝗼𝘁𝗶𝗰𝘀 𝗮𝘁 𝘀𝗰𝗮𝗹𝗲.  1. Searched a 46B compound space using RL to optimize antibacterial activity and solubility simultaneously.  2. Outperformed Monte Carlo and virtual screening baselines, generating 11.6% predicted multi-objective hits vs 0.006% for AI-based screening.  3. Synthesized 79 unique AI-designed compounds; 13 showed in vitro potency (MIC ≤ 8 µg/ml), and 7 were structurally novel.  4. Validated one compound, synthecin, in a mouse MRSA wound model, showing full infection suppression and zero tissue inflammation. Couple thoughts:  • Rather than filtering out high-toxicity candidates post-hoc via ADMET-AI, integrating ClinTox predictions into the RL reward could steer generation away from unsafe chemotypes from the outset.  • Feeding back in vitro MIC and solubility results to continuously retrain the RL value models could sharpen predictions in relevant chemical neighborhoods and expedite SAR optimization, leveraging the strong clustering behavior already observed.  • The current maximal independent set method ensures chemical diversity but can be further enhanced by recent GFlowNet-inspired subset selection algorithms to yield larger, more evenly distributed clusters of candidates. Here's the awesome work: https://lnkd.in/gwVNdtqy Congrats to Kyle Swanson, Gary Liu, Denise Catacutan, Stewart McLellan, Autumn Arnold, Jonathan M. Stokes, James Zou and co! I post my takes on the latest developments in health AI – 𝗰𝗼𝗻𝗻𝗲𝗰𝘁 𝘄𝗶𝘁𝗵 𝗺𝗲 𝘁𝗼 𝘀𝘁𝗮𝘆 𝘂𝗽𝗱𝗮𝘁𝗲𝗱! Also, check out my health AI blog here: https://lnkd.in/g3nrQFxW

AI-enabled drug discovery reaches clinical milestone My piece in @NatureMedicine on exciting progress in our field https://rdcu.be/eugUu Few AI-designed drug candidates have gone beyond in silico benchmarks. Now, a study in Nature Portfolio Medicine reports a successful phase 2a trial of rentosertib, an AI-discovered drug and target combination for idiopathic pulmonary fibrosis What distinguishes this study (in addition to clinical data) is the upstream innovation pipeline This trial marks a turning point: it affirms a potential for AI to do more than generate molecules faster and cheaper; guide discovery, de-risk development and potentially reshape how we develop medicines A pertinent question is: why did this AI-generated drug candidate advance to clinical testing when so many others have not? 🎯 Cross-disease target discovery and 'time-machine' setup: AI models trained on past data predicted therapeutic targets years ahead of traditional methods, pinpointing TNIK as a promising target 🔬 Robust biological validation: Integrated multi-omic analyses, network biology, and extensive literature mining rapidly validated TNIK’s biological relevance for fibrosis ⚙️ Chemistry design: Generative AI models designed molecules targeting novel binding sites, prioritized drug-likeness and synthetic feasibility, and proactively optimized pharmacokinetics and potency from early stages Alex Zhavoronkov Insilico Medicine Harvard Medical School Department of Biomedical Informatics Harvard University Harvard Medical School Harvard Data Science Initiative Kempner Institute at Harvard University Broad Institute of MIT and Harvard

Google DeepMind Spinout Isomorphic Labs Nears Human Trials for AI-Designed Cancer Drugs: 💊The company’s platform is powered by AlphaFold3, the latest iteration of DeepMind’s Nobel-winning AI that predicts protein structures and models drug-target interactions 💊Its lead candidates, including cancer drugs, are currently moving through preclinical development, with human trials expected to begin soon 💊The goal isn’t just one breakthrough drug, but a general-purpose AI engine that can be applied across multiple diseases and modalities 💊The company aims to improve speed, cost, and success rates in drug discovery, reducing pharma’s current 10 percent trial success odds. AlphaFold gives scientists a head start by predicting how well a molecule might bind to a disease-relevant protein target, a key early step in drug design 💊Isomorphic ultimately hopes to turn drug discovery into something closer to design automation: “click a button, get a candidate,” with AI doing the heavy lifting. If successful, it could reshape not just timelines, but how pharma allocates resources and defines early-stage risk 💊Isomorphic has raised $600 million (led by Thrive Capital) and signed major R&D deals with Novartis and Eli Lilly and Company, supporting both external and in-house drug programs #DigitalHealth #AI #Pharma