Kate Beddoe’s Post

It’s an incredible week for #AI and science. Today, the Google Research team and partners shared the most detailed map of the human brain ever created, opening up new possibilities for discovery around neurological disorders, brain structure, and more — a massive feat and a boon for the science community. And earlier this week, Google DeepMind and Isomorphic Labs announced #AlphaFold 3, able to predict all of life's molecules — particularly protein interactions with RNA and DNA — and make it freely-accessible and easy-to-use for the science community through the new AlphaFold Server. These are the AI applications that will move the world forward. Read more about each breakthrough here: 🧠New Images of the Human Brain: https://lnkd.in/d-aRjQVM 🧬AlphaFold 3 and the AlphaFold Server: https://lnkd.in/dEZwBnWU

How does the human brain outperform giant AI models while using a small fraction of the power they require? That's part of what we can hope to unravel by studying the human brain and its synaptic structure. This is a fascinating post about the 10 year quest by the Connectomics team at Google to map the brain. Their work has culminated in a detailed map of 1 cubic mm of the human brain (about 1 millionth of a full brain) which should give us new insights into how brains work. Amazing and inspiring work! https://lnkd.in/gNbqgTfm

Today in partnership with Google DeepMind, we're sharing progress on the jointly developed next generation of AlphaFold. Our new model greatly expands coverage of structure prediction beyond proteins to other key biomolecular classes. This enables our researchers to discover novel ligands mapped to complex biomolecular structures to design better drugs. It also sets a new bar for protein-ligand structure prediction by outperforming the best reported docking methods - without requiring a reference structure. This next milestone for AlphaFold shows AI's potential to continue accelerating biomedical breakthroughs and realise the next era of digital biology. We're making significant progress with Google DeepMind and look forward to sharing more updates: bit.ly/45Qt4yT #AlphaFold #AI #drugdiscovery #machinelearning

An appreciation for how the elegance of mathematics explains everything in minute detail led #BrockU's Yifeng Li, Canada Research Chair in Machine Learning for Biomedical Data Science, to use AI and machine learning to develop and refine drugs to treat cancer and other illnesses. Read more https://loom.ly/KrmEuIA TIPS-SPIIE

🧬 Today, on National DNA Day, we celebrate the milestones of the Human Genome Project's completion in 2003 and the revelation of DNA's double helix structure in 1953. These breakthroughs pave the way for the future of DNA research, powered by AI, ushering in a new era of advancements for biomedicine, personalized medicine and precision healthcare. DNA models like Google's Enformer and NVIDIA's GenSLMs are being developed to improve our understanding of genomic "grammar." These models, along with recent advancements in generative AI, enable DNA analysis to be done faster and more accurately than ever before. So today on National DNA Day, Quantiphi honors these scientific breakthroughs as we continue to drive innovation in healthcare and life sciences, one strand of DNA at a time. #NationalDNADay #Genomics #AI #HealthcareInnovation #PrecisionMedicine #Collaboration #Innovation #Quantiphi #SolvingWhatMatters

So much in awe of the Google Research team in collaboration with Harvard University for studying connectomics to figure out how our brains which consume such little energy are capable of feats that our largest AI systems can't do yet. This connectome describes a piece of human brain tissue the size of half a grain of rice that nonetheless requires 1.4 petabytes (1.4 million gigabytes) to encode — including about 16,000 neurons, 32,000 glia, 8,000 blood vessel cells (for a total of ~57,000 cells) and 150 million synapses. This research may change our understanding of how our brains work. #GoogleResearch #GoogleConnectomics #GoogleDeepMind

The human brain has more than 100 billion cells. How do you find the specific cells involved in Parkinson's and Alzheimers disease? It begins with mapping the brain. Inspiring work from AWS and the Allen Institute to create a comprehensive map of the human brain which can help researchers see through the complexity that has clouded our understanding of brain disease. Utilizing the power of AI, the Brain Knowledge Platform will be open to the scientific community for collaborative breakthroughs. I can't wait to see where this leads us... https://lnkd.in/g_Kd7Fhc

At the bioXcelerate event, Dr. Richard Wyse, Director of Clinical Development at Cure Parkinson's, shared fascinating insights on how AI and biobanks will shape the future of disease research. 🧬 He highlighted AI and machine learning's potential in studying diseases, predicting that biobank-focused research could lead to in-depth analysis of about 500 diseases over the next decade. While most diseases may see incremental improvements, Dr. Wyse suggests some could experience breakthrough 'light bulb moments'. Curious about AI's current impact on research & development? Register for our upcoming webinar to learn more: https://lnkd.in/edBdMtcd #MedicalResearch #AI #Biobanks #FutureOfHealthcare

Google DeepMind’s new AlphaFold AI models all of life’s molecules. Biomodels. Google DeepMind and Isomorphic Labs, a biotech company spun out of DeepMind, recently released an updated version of their AlphaFold algorithm. Using a diffusion model like the one behind OpenAI’s DALL-E, AlphaFold can now model a range of biomolecules—these are the players driving the biology of all living creatures, including proteins, DNA, RNA, and small molecules, such as ligands—and predict how they interact in the body. The algorithm vastly outperformed its peers in some areas (though there’s work to do in others). The companies didn’t openly release the algorithm, as they did its predecessor, but are instead offering an online platform, AlphaFold Server, where biologists can run simulations of biomolecule interactions. #ai #innovation #biomolecule #proteins