How New Discoveries Influence Drug Development

Machine Learning in Preclinical Drug Discovery 🧬💊 Machine learning (ML) is increasingly integrated into preclinical drug discovery, offering promising advancements across hit identification, mechanism-of-action elucidation, and translational investigations. A recent paper in Nature Chemical Biology, "Machine Learning in Preclinical Drug Discovery", provides a thorough analysis of how ML is being utilized to enhance efficiency in early-stage drug development. 🔬 Key Insights from the Paper 1️⃣ Hit Identification & Virtual Screening Traditionally, high-throughput screening (HTS) has been the gold standard for identifying potential drug candidates. However, it is resource-intensive and slow. ML-based virtual screening, powered by deep learning models and molecular featurization techniques, is enabling rapid exploration of chemical libraries far beyond what traditional HTS can achieve. The paper highlights the impact of message-passing neural networks (MPNNs) and Deep Docking as effective methods for prioritizing hit compounds. 2️⃣ Mechanism-of-Action (MOA) Elucidation Understanding how a compound interacts with biological targets is critical for drug development. ML is now playing a pivotal role in MOA elucidation through: AlphaFold and RoseTTAFold: AI-driven protein structure prediction is accelerating target identification and binding site analysis. Generative models: Variational autoencoders (VAEs) and diffusion models are not only aiding in de novo drug design but also helping predict chemical interactions with biological systems. 3️⃣ Translational Investigations & ADMET Predictions Many promising compounds fail in later stages due to poor pharmacokinetics and toxicity profiles. ML is being leveraged to enhance ADMET predictions, improving the likelihood of clinical success. The paper discusses advancements in: Solubility and Lipophilicity Predictions: ML-driven models now outperform traditional log(P) estimations, increasing the reliability of early-stage compound selection. Toxicity Screening: AI-powered tools are improving predictions of hERG binding and organ toxicity, reducing late-stage failures. 🚀 The Future of AI in Drug Discovery While ML is proving to be a game-changer, challenges remain, including data quality, interpretability of AI models, and integration with experimental validation. The paper underscores the importance of open-source datasets, AI transparency, and active learning strategies to enhance model accuracy. 🔗 Read the full paper here: https://lnkd.in/gMtXHrHi AI is reshaping the landscape of drug discovery. As these technologies evolve, collaboration between computational scientists, biologists, and chemists will be critical to unlocking their full potential. #AI #MachineLearning #DrugDiscovery #Pharma #Biotech #ArtificialIntelligence #ComputationalBiology #NatureChemicalBiology

Right now, every CEO is wondering the same thing: “How can artificial intelligence help maximize our impact?”   Delivering on the promise of AI isn’t just good business, it has the potential to help us address some of society’s most pressing challenges. So today, I wanted to offer a closer look at how AI is helping us discover new medicines at Novartis.   The process of identifying a new drug, running patient clinical trials, and bringing it to market takes over a decade. Each new medicine costs on average $2 billion to develop, and we know nearly 9 in 10 of the treatments we work on will fail before they ever reach patients.   A major early step in that process is identifying individual targets in the body that we want to design a drug for. Once we identify that target, which most commonly is a protein, we look for molecules that might address the target’s underlying issue – ultimately those molecule structures form the basis for every successful treatment.   Unlocking the right protein and molecular structures is complex stuff – each step often takes years to get right and our scientists consider billions of potential chemical structures that might lead to effective and safe drug candidates.   AI offers us the chance to accelerate that process. Working with partners at Isomorphic Labs – including members of the Google DeepMind team that were awarded the Nobel Prize this year – we’re now able to do things like model how a protein folds and interacts with the molecules we design. AI models also make it possible for us to analyze different chemical structures simultaneously. It has the potential to add up to significant time savings for our drug development scientists and their work to predict what molecules might treat specific diseases better and faster.   We’re just at the beginning of what this technology can do. As we incorporate AI throughout Novartis’ work, I’m excited to see all the ways it helps us unlock the mysteries of human biology, so we can deliver better medicines that improve and extend patients’ lives.

Pharma R&D has long been one of the most complex and costly innovation challenges in business. Bringing a single drug to market can take over a decade, cost more than $1 billion, and still fail 90% of the time. The process is slow, uncertain, and riddled with scientific and operational bottlenecks—from identifying viable molecules to running large-scale clinical trials and navigating regulatory hurdles.   But our Capgemini AI Futures Lab and our partners at Bayer asked themselves a simple question : 💊 What if new technologies could rewrite this equation? Can #AI rewire the way we develop drugs and reshape the future of #healthcare?   In our latest blog post from the AI Futures Lab, Mark Roberts explores how GenAI is reshaping every phase of pharmaceutical R&D at Bayer—from molecule discovery to clinical trials, manufacturing, and beyond. The convergence of biology, data science, and Generative AI is creating a once-in-a-generation opportunity to rethink how we discover, test, and deliver medicine.   Notably:                                                                                                                                        🔬 "GenAI is already acting as a co-scientist: designing novel molecules, predicting drug interactions, and identifying promising targets with speed and precision never seen before. 👥 In clinical development, AI is enabling smarter patient selection, generating synthetic data, and streamlining regulatory processes—helping reduce trial failure rates and time to approval. 🏭 Beyond the lab, AI is optimizing manufacturing and powering digital therapeutics, ushering in new models of patient care.   So maybe the question is no longer if AI will change pharma—it’s how fast organizations can adapt. 👉 Read our full article here: https://lnkd.in/eebkxpum

Sometimes a breakthrough is about more than just the drug itself—it's about the unique path our technology and team took to develop it. Our RBM39 program at Recursion is a prime example. One of the biggest challenges in drug discovery is that many proteins involved in driving diseases are difficult to drug. Take CDK12, for example. This protein is known to play a key role in certain types of cancer, including those driven by defects in cells’ ability to repair damaged DNA. But, historically, CDK12 has been extremely difficult to target without impacting the closely related CDK13, which can cause serious dose-limiting toxicities. Using our AI-powered maps of biology and Recursion OS platform, we discovered an adjacent protein, RBM39, that appeared to be functionally similar to CDK12, and we hypothesized that designing a drug to target RBM39 might offer a safer alternative. Our scientists then leveraged AI and dry/wet lab validation to design and optimize an RBM39 degrader, now known as REC-1245. In both our maps and preclinical models, REC-1245 mimicked the response of genetically knocking out CDK12 without impacting CDK13. This signaled to us that we were onto something. And notably, we went from target ID to IND enabling studies in under 18 months, more than twice the speed of industry average. We didn’t stop there. We also leveraged our AI platform to better understand which patients might benefit most from REC-1245. Both in-vitro and in-vivo, REC-1245 showed stronger activity in cancer cell lines with high replication stress and poor DNA damage response – both genetic characteristics of tumors that can be tested for in patients. This finding has helped to inform a biomarker-enriched clinical development strategy.   This is a powerful case study of how our platform at Recursion is helping us make informed, evidence-based decisions from end to end throughout the R&D process – from the earliest stages of discovery to molecule design to patient stratification for clinical trials. Of course now is the most important test – how this will perform in trials. REC-1245 is now in a Phase 1/2 study (DAHLIA) and has dosed patients, with enrollment focused on patients with biomarker-selected solid tumors and lymphoma.   Hear more from our scientist Chase Neumann, PhD who’s worked on this program from the beginning, and check out the article in the comments for more details. #DrugDiscovery #TechBio #AI #CancerResearch #PrecisionOncology

🚨 I'm excited to share our latest review, “New approach methodologies for drug discovery,” published in Cell by Cell Press and selected as a Featured Article. For decades, drug discovery has relied heavily on animal models. Yet, with persistently high clinical failure rates, a fundamental question remains: How predictive are animal models of human biology, and are there better alternatives? In this review, we highlight a paradigm shift toward human-centric New Approach Methodologies (NAMs), driven by rapid advances in both policy and technology. On the regulatory front, we discuss major transitions led by agencies such as the FDA (FDA Modernization Acts 1.0 → 2.0 → 3.0) and The National Institutes of Health (stem cell guidelines and the establishment of national organoid initiatives). 🔬 On the technology side, we frame NAMs evolution across three domains: ·        “New” - foundational 2D stem cell–based systems ·        “Newer” - advanced 3D organoid-based models ·        “Newest” - future-facing in silico and AI-driven platforms Across these domains, we highlight emerging therapeutic candidates, cutting-edge models, and translational and clinical applications. We also examine key biological, technical, and regulatory bottlenecks that need to be addressed to enable robust translational adoption, and discuss ongoing clinical efforts and societal considerations for responsible implementation. 🌍 Looking forward. If the past 30 years of drug discovery were shaped by animal models, the next 30 years, animal models will likely transition from a central to a supporting role, following the 3Rs principle: refinement, reduction, and ultimately replacement. Instead, the field will likely be defined by human-centric NAMs, powered by multiscale platforms, multi-omics data, and AI-enabled pipelines. This transformation is not only scientific, but also societal, aligning drug development more closely with human biology while reducing cost, inefficiency, and ethical burden. 👏 Congratulations to an outstanding team: Wenqiang (Eric) Liu, Paul Pang, Catherine Wu and Danilo Tagle from Stanford University School of Medicine, Stanford Cardiovascular Institute, Stanford Department of Medicine, Greenstone Biosciences, National Center for Advancing Translational Sciences (NCATS), The National Institutes of Health. 📄 Please check the full paper here: https://lnkd.in/estaR2cq #NAMs #DrugDiscovery #StemCell #Organoids #AI #PrecisionMedicine #TranslationalScience

I’m thrilled to share that our latest research has been published in Nature Biotechnology! In collaboration with Alán Aspuru-Guzik's lab at the University of Toronto and Insilico Medicine & Alex Zhavoronkov, we’ve demonstrated how quantum computing and AI can enhance the drug discovery process. This approach has allowed us to design new molecules targeting KRAS, a cancer-driving protein long considered “undruggable.” Why is this important? KRAS mutations are present in about 25% of human cancers, including pancreatic, lung, and colorectal cancers. Despite their prevalence, only two FDA-approved drugs, sotorasib (Lumakras) and adagrasib (Krazati), currently target mutant KRAS, offering limited survival benefits. This underscores the urgent need for improved therapies. What makes this approach novel? Traditional drug discovery is time-consuming and resource-intensive. By using quantum computers and generative AI, we’ve been able to simulate and design molecules much faster. Our AI models, trained on a dataset of 1.1 million molecules, helped identify 15 promising candidates for lab testing. Of these, two stood out for their ability to target multiple versions of mutated KRAS in live cells, making them strong candidates for further development. What’s the impact? This hybrid approach could significantly shorten the preclinical phase of drug discovery, making the process faster and more efficient. By leveraging computational methods, we eliminate the need for physical storage of large chemical libraries and robotics for high-throughput screening. Looking ahead While this study is a proof of principle and does not yet demonstrate a significant “quantum advantage” over classical methods, it lays the groundwork for future advancements. As quantum computers become more powerful, they could play an increasingly important role in drug discovery. This is just the beginning. We are already applying this approach to other “undruggable” proteins, aiming to develop treatments for cancers and diseases that currently lack effective therapies. Thank you to many of my lab members, incredible collaborators and everyone who contributed to this exciting project! P.S. In the photo, Alán Aspuru-Guzik (wearing a hat on the left), Alex Zhavoronkov (on the right), and myself in the middle. Christoph Gorgulla Jamie Snider Anna Lyakisheva Zhong Yao Danielle Tahoulas Ardalan Hosseini Petrina Kamya, Ph.D. Alex Aliper Akshat Nigam #CancerResearch #DrugDiscovery #QuantumComputing #AI #KRAS #Biotechnology #Innovation #NatureBiotechnology

Elnora AI is building a 'negative' dataset to improve drug target identification and validation, helping pharmaceutical and biotech companies avoid repeating mistakes and reduce the current 90% failure rate in drug discovery. While pharma companies focus on later-stage development, the crucial work of understanding disease biology, human physiology, and novel drug target discovery is primarily done by academics and emerging techbio companies. These researchers do the heavy lifting in early-stage drug target discovery. However, most of this valuable research never gets published. Even worse, when it does, around 70% of published articles aren't reproducible, with $28 billion spent annually on irreproducible biomedical research in the U.S. alone. And that's not even counting fraudulent papers and paper mills. This crisis has persisted for decades. In 2011, Bayer reported that only 20-25% of 67 investigated studies matched original findings, while in 2012, Amgen found just 11% of landmark cancer studies were replicable. Pharma companies often pick up promising ideas from scientific literature, only to discover that experiments yield different results or protocols don't work as described. It's a systemic problem! Even when they successfully validate experiments and proceed with a drug target, they often miss crucial aspects of the target's biology. This leads to failures in clinical trials due to lack of efficacy (like in a recent Pfizer case) or unexpected toxicity because blocking the target creates unforeseen complications—issues that could have been anticipated with a more complete understanding. Yes, biology is incredibly complex, and it's impossible for any individual to have a complete picture of a drug target. But what if we could tap into collective knowledge? What if we could reach out to hundreds of researchers who have worked on that target but never published their findings because they were "negative" results? That's why we're building a platform to collect unpublished data directly from academic researchers, focusing on detailed lab protocols rather than formal scientific literature. But we're not just taking data—we're creating a system that helps researchers to get their lab experiments working faster and gives the recognition they deserve for sharing their insights. We reward them for contributing negative data that traditional publishing overlooks. By piecing together this puzzle of collective knowledge, we're not just saving time and resources—we're helping save lives. Together.

𝗧𝗵𝗲 𝗺𝗼𝘀𝘁 𝘀𝘂𝗰𝗰𝗲𝘀𝘀𝗳𝘂𝗹 𝗡𝗔𝗠 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗮𝗿𝗲𝗻’𝘁 𝗿𝗲𝗽𝗹𝗮𝗰𝗶𝗻𝗴 𝗮𝗻𝗶𝗺𝗮𝗹 𝗺𝗼𝗱𝗲𝗹𝘀 — 𝘁𝗵𝗲𝘆’𝗿𝗲 𝗿𝗲𝗽𝗹𝗮𝗰𝗶𝗻𝗴 𝘂𝗻𝗰𝗲𝗿𝘁𝗮𝗶𝗻𝘁𝘆. 2025 was a year when New Approach Methodologies (NAMs) dominated press releases. But beyond the publicity, something became very clear to me. In practice, New Approach Methodologies (NAMs) rarely enter pharma R&D as wholesale replacements for existing models. That expectation misunderstands how decisions are actually made in discovery and early development. What NAMs are increasingly good at — and increasingly valued for — is something more specific: 𝘳𝘦𝘥𝘶𝘤𝘪𝘯𝘨 𝘶𝘯𝘤𝘦𝘳𝘵𝘢𝘪𝘯𝘵𝘺 𝘢𝘵 𝘤𝘳𝘪𝘵𝘪𝘤𝘢𝘭 𝘥𝘦𝘤𝘪𝘴𝘪𝘰𝘯 𝘱𝘰𝘪𝘯𝘵𝘴. Whether it’s • flagging hepatotoxicity risk earlier, • stress-testing exposure–response assumptions, or • helping teams walk away from candidates that look promising on paper but fragile in human-relevant systems, the real value lies in 𝘪𝘯𝘤𝘳𝘦𝘢𝘴𝘪𝘯𝘨 𝘤𝘰𝘯𝘧𝘪𝘥𝘦𝘯𝘤𝘦 𝘣𝘦𝘧𝘰𝘳𝘦 𝘥𝘰𝘸𝘯𝘴𝘵𝘳𝘦𝘢𝘮 𝘤𝘰𝘴𝘵𝘴 𝘦𝘹𝘱𝘭𝘰𝘥𝘦. This is why many NAMs gain traction first 𝘪𝘯𝘴𝘪𝘥𝘦 organizations, long before they appear in regulatory-facing packages. They help teams make better go/no-go decisions, prioritize follow-up experiments, and focus resources where they matter most. In doing so, they also contribute to reducing animal studies — by informing earlier decisions not to advance fragile drug candidates into further development. From a development and investment perspective, this shifts the success criteria. Across inventor-stage NAM companies, the strongest ones aren’t trying to solve everything at once. They are: • anchoring themselves to one expensive uncertainty, • generating evidence decision-makers trust, and • integrating into existing workflows rather than challenging them head-on. Over time, 𝘤𝘰𝘯𝘧𝘪𝘥𝘦𝘯𝘤𝘦 𝘤𝘰𝘮𝘱𝘰𝘶𝘯𝘥𝘴. Methods that repeatedly inform good decisions become indispensable — regardless of whether they started as a “replacement” for anything. Perhaps that’s the real adoption pathway for NAMs: not as short-term alternatives to legacy models, but as 𝘲𝘶𝘪𝘦𝘵 𝘪𝘯𝘧𝘳𝘢𝘴𝘵𝘳𝘶𝘤𝘵𝘶𝘳𝘦 𝘧𝘰𝘳 𝘣𝘦𝘵𝘵𝘦𝘳 𝘥𝘦𝘤𝘪𝘴𝘪𝘰𝘯𝘴. #NAMs #PharmaR&D #DrugDevelopment #Preclinical #Biotech

A fascinating Perspective published in Nature Chemical Biology (Nature Portfolio) by Tim Stachowski, PhD and Marcus Fischer sheds light on something rarely discussed outside structural biology: 𝗨𝗻𝗱𝗲𝗿 𝗰𝗿𝘆𝗼𝗴𝗲𝗻𝗶𝗰 𝗰𝗼𝗻𝗱𝗶𝘁𝗶𝗼𝗻𝘀 𝘂𝘀𝗲𝗱 𝗶𝗻 𝗺𝗮𝗻𝘆 𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗮𝗹 𝗺𝗲𝘁𝗵𝗼𝗱𝘀, 𝗽𝗿𝗼𝘁𝗲𝗶𝗻𝘀 𝗰𝗮𝗻 𝗹𝗼𝘀𝗲 𝗮𝘀𝗽𝗲𝗰𝘁𝘀 𝗼𝗳 𝘁𝗵𝗲𝗶𝗿 𝗻𝗮𝘁𝘂𝗿𝗮𝗹 𝗰𝗼𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗺𝗼𝘁𝗶𝗼𝗻. As a result, the structure captured under cryo conditions isn’t always the conformation a drug encounters in real biological environments. Recent studies show that when proteins are examined at higher, near-physiological temperatures, additional conformations appear – and some of these play a meaningful role in how molecules bind. These dynamic states can open the door to entirely new design ideas. 𝗙𝗼𝗿 𝘁𝗵𝗼𝘀𝗲 𝗼𝗳 𝘂𝘀 𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝗶𝗻 𝗰𝗼𝗺𝗽𝘂𝘁𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗮𝗻𝗱 𝗔𝗜-𝗲𝗻𝗮𝗯𝗹𝗲𝗱 𝗱𝗲𝘀𝗶𝗴𝗻, 𝘁𝗵𝗲 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆 𝗶𝘀 𝘀𝘁𝗿𝗮𝗶𝗴𝗵𝘁𝗳𝗼𝗿𝘄𝗮𝗿𝗱: the closer the structural model is to real biology, the more reliable the predictions. Importantly, this perspective builds on the foundations laid by cryo-based structure determination – now the main engine of protein structure discovery – while highlighting why capturing protein dynamics is increasingly critical for drug design. This is why at Pepticom Ltd. we focus on flexible, dynamic targets rather than single frozen conformations – an approach very much supported by the direction outlined in this article. Worth reading for anyone following how protein dynamics will shape the next generation of drug discovery > https://lnkd.in/er3eZFB9 #AI #Peptides #DrugDesign #DrugDiscovery

Scientists have discovered that changing just two atoms in the structure of LSD may remove its hallucinogenic effects while preserving its therapeutic potential. This modified compound, often described as a nonhallucinogenic psychedelic, showed strong ability to activate the same healing pathways in the brain without causing altered perception. Early research suggests that the new molecule may help treat depression, anxiety, trauma related disorders, and other mental health conditions with fewer risks than traditional psychedelic therapy. The idea is to keep the benefits of the drug while avoiding the intense experiences that make it unsuitable for many patients. In laboratory studies, the altered molecule activated serotonin receptors linked to mood regulation and emotional processing. These receptors help guide neuroplasticity, the brain’s ability to adapt and reorganize. Many mental health conditions are connected to reduced plasticity, leaving patients stuck in harmful thought patterns. By stimulating these pathways without inducing hallucinations, the new molecule may offer a safer and more accessible treatment option. Researchers report that animals treated with the modified compound showed reduced anxiety behaviors and improved stress resilience. The breakthrough may also help people who cannot undergo psychedelic assisted therapy because of medical conditions, personal concerns, or sensitivity to hallucinogenic effects. A treatment that works without supervision in a clinical session could give millions wider access to care. Scientists are now studying how long the therapeutic effects last, how the drug behaves in larger doses, and whether it interacts safely with other medications. Human trials will be required before the therapy becomes available, but the progress is promising. This discovery represents a growing shift in mental health research. Instead of relying only on traditional antidepressants, scientists are exploring how precise molecular changes can influence deep emotional circuits. If future studies confirm these findings, the modified LSD compound may reshape how depression and anxiety are treated and offer hope to people who have not responded to current options. #BetterU #PsychedelicTherapy #Science #Health #Innovation