Samuel Scarpino’s Post

At Johnson & Johnson, we’re leveraging the power of technology to unlock cutting-edge science and deliver potentially game-changing innovation for patients around the world.      We’re excited to collaborate with BigHat Biosciences to use artificial intelligence (AI) and machine learning (ML) to expand our capabilities in antibody discovery and design. By combining our team’s expertise in drug discovery, clinical development and data science with BigHat’s AI/ML technologies, integrated with a high-speed wet lab, our aim is to more quickly engineer antibodies with better biophysical properties against difficult-to-treat diseases in neuroscience.     Learn more about this collaboration and its potential impact: https://bit.ly/3Qg7Bui     #DrugDiscovery #DataScience #ArtificialIntelligence #MachineLearning  

🔬✨ 𝐀𝐈 𝐢𝐧 𝐃𝐫𝐮𝐠 𝐃𝐢𝐬𝐜𝐨𝐯𝐞𝐫𝐲: Pioneering the Future! 🚀 - Accelerated Discovery: With AI, the timeline from initial screening to potential drug candidate is dramatically reduced, getting solutions to patients faster! ⏱️💊 - Enhanced Precision: Machine learning algorithms enable precise targeting of molecules, minimizing the risk of side effects and optimizing efficacy. 🎯🔍 - Cost-Effective Solutions: AI significantly cuts R&D costs, freeing up resources for more innovative exploration and investment! 💡💰 - Data-Driven Insights: From genomic data to clinical outcomes, AI crunches vast datasets, uncovering hidden patterns and novel therapeutic pathways. 📊🔬 Curious to keep up with cutting-edge research? Check out Sciqst for your next biomedical literature review! 📚➡️ https://www.sciqst.com #AIDrugDiscovery #BiotechRevolution #Medicine #InnovationInScience

Reposting to increase the number of people who see this and could answer my question: Can someone share the peer reviewed source of work described in the Ted Talk below? If it’s in the Ted talk, can you tell me where approximately? I’d like to read the paper(s) and understand better what “predicted the structures of all 200 million proteins known to nature means”. My knowledge of protein structure prediction is limited primarily to x-ray crystallography, NMR and homology models based on those structural models that are stored in the protein database (PDB). I’d like to have a sense of how accurate the structures are predicted to be and how that accuracy was determined. My major concern is that predicting the protein structures of every proteins that is even remotely related to a solved structure is exciting but not necessarily new given that it was frequently done 18+ years ago. If the prediction is accurate for proteins that have no similarity to previously solved proteins, how do we know it’s right without first solving some of those sequences. Even if you used a subset of the PDB to train a model and another subset to test those predictions, I would think there’d be a big level of uncertainty of the predictions that are for protein with no practical similarity in the PDB. Update: Thanks to Daniel Hickmore for the DM it got me to the Nature paper that has been accepted but is not yet published. I'm reading it now to see if my questions get answered: https://lnkd.in/ei3pyCax

🚀 Exciting Breakthrough in AI-Powered Drug Repurposing! 🧬💊 Harvard Medical School researchers have developed TxGNN, a game-changing AI model that could revolutionize treatment options for rare and neglected diseases. With only 5-7% of 7,000+ rare diseases having FDA-approved treatments, this innovation couldn't come at a better time! Key highlights: ✅ Analyzed nearly 8,000 existing medicines for 17,080 diseases ✅ Outperformed leading AI models by 50% in identifying drug candidates ✅ 35% more accurate in predicting contraindications ✅ Uses a pre-trained graph neural network for zero-shot inference on new diseases What sets TxGNN apart? • Employs geometric deep learning and metric learning to transfer knowledge between diseases • Incorporates a vast biological knowledge graph for accurate predictions • Provides explainable multi-hop paths for better understanding of predictions The best part? TxGNN is freely available to the scientific community! Researchers are already collaborating with rare disease foundations to explore new treatment options. While additional evaluation is needed before clinical application, TxGNN represents a major leap forward in addressing health disparities and offering hope to patients with rare and untreated conditions. Image source: Nature articles/s41591-024-03233-x Follow Leon Zheng to join the unique journey of AI! #AI #DrugDiscovery #RareDiseases #HealthcareInnovation #MedicalResearch

AI just predicted the structure of 200 million proteins in less than a year: In a recent TED talk, Demis Hassabis, co-founder and CEO of Google DeepMind, highlighted the protein folding problem. Over the past 40 years, experimental biologists have pieced together ~150,000 protein structures. It takes one PhD student their whole PhD (4-5 years) to uncover one structure. But there are 200 million proteins known to nature. This would take forever. AlphaFold predicted the structures of all 200 million proteins known to science in less than a year. A task that would have otherwise taken PhD students billions of years. The predictions are made within the width of an atom—a level of accuracy that allows biologists to use them for disease understanding and drug design. This will fundamentally transform drug discovery from years to months. If you enjoy insights like this, follow me Alex Banks for more on AI. P.S. Join 40,000+ others in my free newsletter to stay on the cutting edge of AI: https://lnkd.in/dPNxW8-Y

🧬 Breakthrough in AI-Driven Drug Discovery: AlphaFold 3 Unveiled! 🚀 Google DeepMind and Isomorphic Labs have introduced AlphaFold 3, a game-changing AI that models DNA, RNA, and facilitates drug development. Here's why it's revolutionary: 🔬 50% more accurate than traditional methods in predicting biomolecular interactions 🧪 No pre-existing structural data needed 🏥 Accelerates research in medicine, agriculture, and materials science 💊 Streamlines drug development processes 🖥️ Accessible via AlphaFold Server for scientists worldwide At Allmatics, we're thrilled to see AI advancements in #healthtech. AlphaFold 3 represents a quantum leap in the industry, potentially saving millions of research years and drastically reducing costs in structural biology. This evolution in health-related #AI is incredibly inspiring. We're excited to continue our #development at the intersection of AI/ML/DS and Healthtech. Want to learn more about AlphaFold 3 and our insights on AI in #healthcare? Check out the links below! Google DeepMind blog: https://lnkd.in/etiXfCw8 AI + Healthcare Integration: https://lnkd.in/eCnkZgdK #AIinHealthcare #DrugDiscovery #AlphaFold3 #GoogleDeepMind

Our #AI innovations, including Target and Lead Identification Suite and Multiomics Suite, are transforming drug discovery and precision medicine. Learn how researchers are harnessing our AI and high-performance computing technologies today and how we foresee #generativeAI advancing the field of life sciences ↓

𝗔𝗜 𝗕𝗿𝗲𝗮𝗸𝘁𝗵𝗿𝗼𝘂𝗴𝗵: 𝗣𝗲𝗽𝗙𝗹𝗼𝘄 𝗥𝗲𝗱𝗲𝗳𝗶𝗻𝗲𝘀 𝗣𝗲𝗽𝘁𝗶𝗱𝗲 𝗦𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻 🧪💻 The 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 𝗼𝗳 𝗧𝗼𝗿𝗼𝗻𝘁𝗼 has just raised the bar in computational biology with #PepFlow, a new #AI model in peptide structure prediction. Here are the key takeaways. 🔷 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 -Uses deep learning inspired by 𝗕𝗼𝗹𝘁𝘇𝗺𝗮𝗻𝗻 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗼𝗿𝘀. -Predicts entire "energy landscapes" of peptides, not just single structures. -Can handle complex peptides, including circular ones formed through macrocyclization. 🔷 𝗖𝗼𝗺𝗽𝗮𝗿𝗶𝘀𝗼𝗻 𝘁𝗼 𝗔𝗹𝗽𝗵𝗮𝗙𝗼𝗹𝗱2 -While #AlphaFold2 revolutionized protein structure prediction, it struggled with flexible peptides. -#PepFlow addresses this limitation, outperforming AlphaFold2 for peptide structures. 🔷 𝗜𝗺𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗳𝗼𝗿 𝗗𝗿𝘂𝗴 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆 -Peptide-based drugs (like GLP1 analogues) are increasingly important in medicine. -PepFlow could accelerate the design of peptides with specific therapeutic properties. -Potential for more targeted, less toxic, and easier-to-produce medications. 🔷 𝗙𝘂𝘁𝘂𝗿𝗲 𝗜𝗺𝗽𝗮𝗰𝘁 -Demonstrates AI's growing capability to model complex biological systems. -Could lead to new insights in fundamental biology and biochemistry. -Shows how specialized AI models can outperform generalist models in niche areas. This breakthrough highlights the synergy between AI and scientific research. As we develop more AI tools, we're unlocking new capabilities in fields from drug discovery to materials science. #AI #ComputationalBiology #DrugDiscovery #MachineLearning

Revolutionizing biomedical research! Rejuve.Bio’s new AI doesn’t just generate hypotheses—it explains them. Discover how our platform is uncovering hidden drug targets and bringing life-saving therapies to market faster. Watch the demo and see how explainable AI is transforming healthcare. https://lnkd.in/gF_RhzGr #AI #Biotech #Longevity #rejuvebio

𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗶𝗻𝗴 𝗣𝗮𝗿𝗸𝗶𝗻𝘀𝗼𝗻’𝘀 𝗗𝗶𝘀𝗲𝗮𝘀𝗲 𝗧𝗿𝗲𝗮𝘁𝗺𝗲𝗻𝘁 𝘄𝗶𝘁𝗵 𝗔𝗜-𝗗𝗿𝗶𝘃𝗲𝗻 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 Researchers at the University of Cambridge have harnessed the power of artificial intelligence to dramatically speed up the search for treatments for Parkinson’s disease. Utilizing AI-based techniques, the team identified compounds that block the harmful aggregation of alpha-synuclein, a protein linked to the condition. 𝗞𝗲𝘆 𝗔𝗰𝗵𝗶𝗲𝘃𝗲𝗺𝗲𝗻𝘁𝘀: 1. Rapid screening of millions of compounds using machine learning. 2. Identification of five highly potent compounds for further exploration. 3. Reduction of initial screening costs by a thousand-fold. 4. Acceleration of the screening process by ten times. These advancements, reported in Nature Chemical Biology, highlight the potential of AI in transforming drug discovery, making the development of treatments faster and more cost-effective. This could significantly shorten the time it takes for potential treatments to reach patients. 🧐 𝗪𝗵𝘆 𝗜𝘁 𝗠𝗮𝘁𝘁𝗲𝗿𝘀: Parkinson's is the fastest-growing neurological condition globally, affecting millions. Traditional drug discovery processes are costly and time-consuming. This innovation not only speeds up research but also opens new doors for multiple drug discovery programs simultaneously. Let’s celebrate the strides AI is making in healthcare and its potential to change lives. Kudos to the Cambridge team for leading the way in AI-powered medical breakthroughs! At Medvolt, we are also harnessing the power of generative AI, alongside other large language models (LLMs) and deep learning technologies, through our innovative platform 𝐌𝐞𝐝𝐆𝐫𝐚𝐩𝐡. 𝐌𝐞𝐝𝐆𝐫𝐚𝐩𝐡 is specifically designed to pioneer novel drug discoveries, contributing to cutting-edge developments in the field. Our commitment to integrating advanced AI tools reflects our dedication to leading the charge in transformative healthcare solutions. Contact us today for the free demo, here is our email: contact@medvolt.ai #ai #drugdiscovery #drug #data #Parkinsons #oncology #cancer #llm #genai #ml