AI in Molecular Prediction

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.

We are proud to present our latest paper on physics-informed AI for drug design appearing in PNAS special issue on machine learning in chemistry . Standard data-driven AI does not work well on examples that are significantly different from training data. This can result in unphysical predictions that are clearly wrong. To limit this type of unphysical result in the realm of drug design we introduced a new machine learning model called NucleusDiff, which incorporates a simple physical idea into its training, greatly improving the algorithm's performance. NucleusDiff ensures that atoms stay at an appropriate distance from one another, accounting for physical concepts such as repellant forces that prevent atoms from overlapping or colliding. Rather than accounting for the distance between every single pair of atoms in a molecule, which would be expensive, NucleusDiff estimates a manifold, and on that manifold, it then establishes main anchoring points to watch, making sure that the atoms never get too close to one another. We predicted binding affinities of a newer molecule that was not included in the training dataset: the COVID-19 therapeutic target 3CL protease. NucleusDiff showed increased accuracy and a reduction of atomic collisions by up to two-thirds as compared to other leading models.

"A.I. hallucinations... are dreaming up riots of unrealities that help scientists track cancer, design drugs, invent medical devices, uncover weather phenomena and even win the Nobel Prize." A nice New York Times article "How Hallucinatory A.I. Helps Science Dream Up Big Breakthroughs" delves into the value of "hallucinations" in scientific advances. (Gift link in comments). Examples in the article include: 🌟 Nobel Prize recognition for "De novo protein design" David Baker’s groundbreaking work at the University of Washington has redefined what’s possible in protein engineering. Using AI hallucinations, his lab designed entirely new proteins from scratch—an achievement once considered "almost impossible." These proteins, numbering over 10 million, include innovations like cancer treatments and tools for combating viral infections. Baker’s work earned him the 2023 Nobel Prize in Chemistry. 🏥 Medical innovation with AI-designed catheters Anima Anandkumar and her team developed a novel catheter design using AI hallucinations to combat a major global health issue: urinary tract infections. Their model generated thousands of possible geometries before selecting one featuring sawtooth-like spikes lining the inner walls. These spikes prevent bacteria from adhering and traveling upstream to the bladder, drastically reducing bacterial contamination. The device is currently under discussion for commercialization. 💊 Accelerated drug discovery MIT professor James J. Collins is using AI to transform antibiotic discovery. By prompting models to dream up completely new molecular structures, his team can quickly identify promising drug candidates. This process, which used to take years, now takes just days, speeding up the fight against drug-resistant bacteria. Collins highlights hallucinations as a tool for sparking creativity in molecular design, a critical area for global health. 🌪️ Advances in weather forecasting Amy McGovern’s work at the University of Oklahoma shows how A.I. hallucinations can improve weather predictions. By generating thousands of probabilistic forecast variations, AI helps uncover hidden factors driving extreme weather events like heat waves. McGovern describes these AI outputs as invaluable for spotting unexpected patterns in the atmosphere. 🖼️ Sharpening medical imaging At Memorial Sloan Kettering Cancer Center, Harini Veeraraghavan has used AI hallucinations to improve medical imaging. By applying the technology to sharpen blurry MRI scans, her team enhances diagnostic accuracy. Their work, described as “hallucinated MRI,” has the potential to change how radiologists interpret scans, especially when clarity is crucial for finding abnormalities. The lesson: hallucinations are a feature, not a bug, if we understand their nature and use AI outputs appropriately.

For 50 years, a key protein behind heart disease, among the leading cause of death worldwide remained a scientific mystery. It was too large and complex for traditional methods; its structure was invisible to us. Now, researchers have combined cryo-electron microscopy with DeepMind's AlphaFold to reveal the atomic structure of that protein: apoB100, the very scaffold of "bad cholesterol." This marks a deeper shift in how we approach science.  When we can see biology at this level of detail, healthcare moves from managing symptoms to engineering interventions at the molecular root. AI starts to function as a new kind of microscope, one that reveals the invisible machinery of life and allows entirely new questions to be asked. This is the kind of progress that matters.     AI as an instrument for understanding, precision, and prevention. It’s a glimpse into a future where compute and science converge to tackle humanity’s hardest health challenges at their source.    Read the full story: https://lnkd.in/gbum2dKu #AIInHealthCare #AIForGood

I’m often asked where I see AI make a tangible, real impact in the world today. To that, I answer with #AlphaFold, the revolutionary AI model from Google DeepMind, that is able to predict the structure of a protein simply from its amino acid sequence. 5 years ago, AlphaFold solved the 50-year grand challenge of protein folding, followed by the equally meaningful decision to make 200 million protein structures freely available to the scientific community. Since then, Demis Hassabis and John Jumper have been recognized with a Nobel Prize for their work on AlphaFold, and we see over 3.3 million users of it globally, with more than a third of users right here in Asia-Pacific. Here is just a snapshot of those applications: 🔬 Dr. Su Datt Lam at the National University of Malaysia (UKM) is learning more about Melioidosis to better fight the silent killer. 🧬 Researchers Lim Jackwee lim and Yinxia Chao at Singapore’s A*STAR - Agency for Science, Technology and Research and National Neuroscience Institute (NNI) are visualizing proteins linked to Parkinson’s. 🔍 Professor Ji-Joon Song’s team at the Korea Advanced Institute of Science and Technology lead to cancer and other diseases. 🪢Dr. Danny Hsu at Academia Sinica, Taiwan is advancing our understanding of exceptionally complex protein “knots”. ♨️ Dr. Syun-ichi Urayama’s team is uncovering new evolutionary insights from microbes in Japan’s hot springs! Listen to one of their stories below, and read more about all of them here: https://lnkd.in/d7wyACpK #GoogleDeepMind #AIforGood 

AI adding value: Stanford University researchers who discovered a new molecule that rivals Ozempic, say that the "study would not have been possible without the use of artificial intelligence to weed through dozens of proteins in a class called prohormones." Professor Katrin J S. PhD said “The algorithm was absolutely key to our findings.” From an article on the discovery: "Instead of manually isolating proteins and peptides from tissues and using techniques like mass spectrometry to identify hundreds of thousands of peptides, the researchers designed a computer algorithm they named Peptide Predictor to identify typical prohormone convertase cleavage sites in all 20,000 human protein-coding genes. They then focused on genes that encode proteins that are secreted outside the cell — a key characteristic of hormones — and that have four or more possible cleavage sites. Doing so narrowed down the search to 373 prohormones, a manageable number to screen for their biological effects."

Five years ago, AlphaFold solved the protein structure prediction problem at CASP14, cracking a 50-year grand challenge in biology. It has been an absolute honour and privilege to have been part of this journey alongside Demis and John. Over 3 million researchers across 190 countries have since used AlphaFold to predict the structure of more than 200 million proteins. The impact spans from revealing apoB100's structure, advancing heart disease research, to supporting endangered honeybee conservation in Europe. Protein structure prediction was the root node problem in structural biology. By solving it, we opened up entirely new avenues for discovery. What AlphaFold demonstrated is that AI can accelerate scientific progress when applied to the right foundational challenges. We've since expanded this approach across biology. AlphaMissense and AlphaGenome are helping researchers understand genetic mutations and disease. AlphaProteo is designing new protein binders for targets in cancer and diabetes. We're applying similar thinking to challenges in fusion energy, materials discovery and climate science. Today, we're sharing The Thinking Game, following our team through the journey that made AlphaFold possible. To understand more about AlphaFold's impact, see the blog here: https://lnkd.in/eiPSAeKc #AlphaFold #AIforScience

AI is incredibly good at optimizing. It can analyze large datasets, identify patterns, and recommend the “best” compound within a given chemical series. That’s powerful — but it’s not enough. Because optimization is not innovation. One of the biggest reasons AI hasn’t yet delivered a wave of breakthrough drugs is this: the models we’re using are largely trained to refine what's already known, not to imagine what’s possible. Most AI systems in drug discovery work by learning from historical data — known compounds, known targets, known properties. From there, they generate new molecules that are similar, but slightly improved. They might bind a little tighter. Be a little less toxic. Have slightly better solubility. This is incremental progress. And it’s valuable, especially in lead optimization. But when the industry talks about "AI discovering new drugs," we often imagine something more revolutionary: first-in-class molecules, targeting undruggable proteins, opening up new biology. That kind of leap requires creative generalization. It demands the ability to infer beyond the data it was trained on. But generative models are, by design, bound by their training data. If the model has never “seen” a class of molecules, or if a protein target lacks good structural data, the system has little foundation on which to innovate. And here’s the paradox: the more robust and curated your training data, the more your model becomes anchored in the past. So, how do we break this loop? Instead of asking AI to generate the “best” compound, we should ask it to generate diverse hypotheses — especially ones that break outside conventional chemical space. Then we need wet-lab systems and organizational cultures that are willing to test bold ideas, not just safe bets. AI’s real potential in drug discovery won’t be realized through faster optimization. It will come from enabling smarter exploration — of targets, of modalities, of mechanisms we haven’t yet charted. #drugdiscovery #AIinHealthcare #biotech #machinelearning #lifesciences #computationalchemistry #generativemodels #pharma #innovation #futureofmedicine

We are nearing the limits of the known antibiotic universe. For decades, progress has largely meant revisiting familiar molecules, even as resistance continues to outpace discovery. A recent effort from Massachusetts Institute of Technology changes the nature of the search itself. Instead of screening what already exists, researchers used generative AI to design tens of millions of hypothetical compounds that have never been synthesized or cataloged before. This is not deeper exploration of known space, but the creation of entirely new chemical territory. The AI generated molecules from first principles, guided by rules of efficacy and synthesizability. Several candidates that emerged are structurally unlike existing antibiotics and appear to act through a more fundamental mechanism: disrupting bacterial cell membranes. That distinction matters. Resistance often develops against drugs targeting specific internal proteins, but compromising the membrane is a broader, harder-to-defend strategy. In early studies, one AI-designed compound proved effective against drug-resistant gonorrhea by targeting a novel membrane-related protein, while another cleared MRSA infections in animal models, operating outside known antibiotic classes. The deeper shift here is conceptual. Generative models expand discovery beyond what can be searched or screened, into what can be designed. At a time when antimicrobial resistance is a growing global threat and the traditional pipeline is stagnant, this exploration-first approach offers a credible path forward. The next chapter of antibiotic development may depend less on rediscovery, and more on invention. #ArtificialIntelligence #GenerativeAI #DrugDiscovery #AntibioticResistance #AntimicrobialResistance #ComputationalBiology #AIinHealthcare #Biotech #LifeSciences #ScientificInnovation

AI just ran its own multidisciplinary tumor board. And nailed the diagnosis + treatment. This was a full-stack oncology reasoning engine—pulling from imaging, pathology, genomics, guidelines, and literature in real time. A new paper in Nature Cancer describes how researchers built a GPT-4-powered multitool agent that: • Interprets CT & MRI scans with MedSAM • Identifies KRAS, BRAF, MSI status from histology • Calculates tumor growth over time • Searches PubMed + OncoKB • And synthesizes everything into a cited, evidence-based treatment plan In short: it acts like a multidisciplinary team. Results : • Accuracy jumped from 30% (GPT-4 alone) to 87% • Correct treatment plans in 91% of complex cases • Every conclusion backed by a verifiable citation This is bigger than oncology. Any field that relies on multi-modal data and cross-domain reasoning—like my field of GI ( GI + Mental Health+ Nutrition + Excercise ) could benefit from this collaborative AI architecture. Despite the visual, it doesn’t replace the human team—it augments it. Providers still decide. But now, they do it faster, with more context, and less cognitive fatigue. #AI #HealthcareonLinkedin #Healthcare #Cancer