Data Care provides an HPC platform for Drug Discovery, ML Training and large scale simulations. 1000's of nodes based on NVIDIA A4000 with Ampere architecture with more than 6000 cuda cores.
HPC, Drug Discovery, ML Training Cloud, Computational Chemistry, Computational Biology, BioInformatics, AI Infrastructure, Medical Informatics, Clinical Trials, Generative AI, Health Care AI, and AI Research
Investment Partner, Bio + Health @ Andreessen Horowitz
As AI masters the language of biology, the ways we interpret science, craft hypotheses, design cures, and fuel the broader bio-economy are poised for radical improvement. Simply put, the recent explosion of models across the drug discovery lifecycle promises to usher in a new era of medicine.
Today, my a16z Bio + Health colleagues, Vijay Pande, PhD, Becky Pferdehirt, PhD, Zak Doric, PhD, and I published some of the most pressing “Jobs to be done” for AI in the life sciences.
Get the full insights at the link below!
AI and Synthetic Biology are Set to Revolutionize The World!
💰 The global synthetic biology market was estimated to be worth $11.4 billion in 2022 and is projected to reach $35.7 billion by 2027.
🧬 Synthetic biology, a field that combines biology and artificial intelligence, is revolutionizing various sectors including medical science, pharmaceuticals, food science, agriculture, and climate change research.
💻 The practice involves reprogramming existing proteins or biological material to achieve new functions.
🚀 The rise of cloud and distributed computing is accelerating this effort, enabling rapid genetic and DNA sequencing.
🧪 This technology allows scientists to manipulate and redesign cells to drive specific outcomes, from biofuels to disease-resistant plants.
🏭 Synthetic biology companies see opportunities for AI throughout the product-development life cycle, from initial design to testing, and even in scaling up manufacturing.
Arzeda
Full article: https://lnkd.in/eUUfvuHm#SyntheticBiology#ArtificialIntelligence#BioTech#FutureOfScience#AIinBiology#RevolutionInResearch#DNASequencing#CloudComputing
Cracking the Code: How AI Predicts Drug Potency
AI Approach: Utilizes Graph Neural Networks (GNNs) for drug potency prediction.
Learning Process: GNNs remember known data, and sparingly use learned chemical interactions.
Methodology: Evaluated six GNN architectures using EdgeSHAPer method.
Focus Area: GNNs primarily concentrate on ligands, less on protein-drug interactions.
Challenge: Overreliance on chemically similar molecules raises concerns about prediction accuracy.
Potential Improvement: Some GNN models show promise for enhancement via modified representations and training techniques.
Importance: Highlights the need for explaining complex AI predictions in pharmaceutical research.
Overall Theme: Unveiling the mechanisms behind AI's drug potency predictions.
Full Article: https://lnkd.in/dZxRHHbe#AIPharmaInsights#DrugDiscovery#GNNResearch#AIInMedicine#MachineLearningHealth#CompoundPotency#PharmaTech#EdgeSHAPerAnalysis#Bioinformatics#FutureOfMedicine
Cost effective and portable BP screening coupled with cloud based simple yet scalable EMR can be a very useful tool for public health projects. My team is happy to conduct trials on this as part of vision screening camps.
Perhaps the same device can as well be used for other vital signs capture as well with minimal tweaking for pulse ox, HR… and cardiac rhythm abnormalities like Afib.
While AliveCor, Apple iWatch proved and have gotten the FDA approval for A. Fib rhythm abnormality detection using single lead EKG data, I see no issue the same being achieved with waveform detection tools.
Among R-R interval variability and absence of p wave in EKG, second factor is not as significant as the first one. This would be easily detectable with wave form irregularities due to changes in cardiac filling and subsequent changes to stroke volume and wave form amplitude changes. Only question is that of sensitivity and accuracy. Most likely one might need a bit longer measurement ( 10-15 seconds instead of 5-6 seconds ) , but all of this can be tweaked with a bit of model training.
If the farm factor of this sensor be adapted to Oura Ring type of farm factor, one key invention could be diagnosing "pre-hypertension" by looking at nocturnal variation of blood pressure and identifying "non-dippers" years before they formally become hypertensives.
Tobin Lim, MD, ABIM, Dhanunjaya (DJ) Lakkireddy MD, MBA, Senthil Nachimuthu, Raghu B., Bhava Reddy, David Albert, Yashwanth Kotha, Chaitanya Mamillapalli, MD, MRCP, FAPCR, FACE, Rajesh Ravuri, Boaz Rosen, Ajit Kesani, Vijayabhaskar Reddy Kandula, Manu Dasari, Rajesh Vukkala, Rajasekar Bhoda, MD, MBA, Dr. Sowmya gowda, Devaraj M., Branden D. Rosenhan, M.D.
Digital Health | Interactions for Telehealth Applications | Human Factors | UX Research & Design | Deep Learning| PhD Candidate at UC San Diego
Today the Digital Health Technologies Lab was present at the Center for Wearable Sensors Research Summit at UC San Diego Jacobs School of Engineering! Me, Nancy and Ava Fascetti got to share with industry, faculty and students the amazing projects our lab is working on. Nancy (Yujia Liu) brought home the best pitch award. Congrats Nancy! It was great to share our work and to find possible research collaborations within the CWS community.
Accurate proteome-wide missense variant effect prediction with AlphaMissense
With the recent FDA approval of CRISPR for sickle cell and very likely for thalassemia.. perhaps researching single nucleotide substitute leading to diseases across the human genome is equally important.
Read here: https://lnkd.in/dNJMJsbU
Cutting-edge AI meets molecular sciences at NeurIPS 2023! 🌐🔬
Small-Molecule Discovery:
# DL, SMILES, and graph methods advance drug discovery.
# Attention-based pooling proves effective.
# Generative models aid molecule design using learned reactions.
Protein Structure:
> AlphaFold predictions match with some discrepancies.
> pLDDT scores gauge prediction confidence.
> Combining AlphaFold2 and cryo-EM enhances accuracy.
Top 3 NeurIPS Posters:
> OpenProteinSet: Scale training for structural biology.
> ProteinInvBench: Benchmarks Protein Inverse Folding.
> Protein Design with Guided Discrete Diffusion.
https://lnkd.in/ddQC4S6y
