Congratulations to Drs. David Baker, Demis Hassabis, and John Jumper for winning The Nobel Prize in Chemistry! This marks the second Nobel Prize for AI in 2 days! At 1910, our work in large molecule AI drug discovery began with AI-driven protein structure prediction. In 2018, tools like Rosetta were able to predict the N-terminal to C-terminal backbone structure of proteins but sometimes, not with high enough resolution to predict the conformations of the unique side chain attached to each amino acid. Protein structures are defined by hydropathic interaction (HINT) networks on multiple scales. These networks are hydrophobic as well as electrostatic/hydrogen bonding, and include favorable and unfavorable elements. Three-dimensional (3D) hydropathic interaction maps calculated by HINT encode these networks on a residue-by-residue basis. Importantly, these HINT maps can be clustered – reducing thousands of residue environments into ~18 maps on average – each with a unique side chain conformation and set of interactions. Through a 2019 collaboration with Dr. Glen Kellogg of Virginia Commonwealth University, 1910 developed a methodology for processing all 20 amino acids, thereby generating 12,000,000 HINT maps, yielding > 4,000 unique backbone maps and >10,000 unique side chain maps for about 800 total chess squares. This dataset was an unprecedented collection of information-rich 3D maps that encoded the hydropathic interaction environments favored by each amino acid residue. We then built AI and ML models to leverage our proprietary HINT map dataset for a variety of applications including side chain prediction, homology modeling, 3D protein structure prediction, and de novo protein design. With AlphaFold2, predicting both the structure of protein backbones and amino acid side chain conformations became possible, thereby solving a holy grail problem that was intractable for decades. While our work in large molecule AI drug discovery had roots in the prediction of protein side chain conformations, it has since expanded to include de novo design and optimization of large molecule therapeutics such as traditional monoclonal antibodies, Fabs, VHHs, antibody-drug conjugates, etc. Congratulations to Drs. Baker, Hassabis and Jumper for their well-deserved Nobel Prize, and for showing yet again, the tremendous promise of AI to revolutionize drug discovery!
1910 Genetics’ Post
More Relevant Posts
-
New #AI system designs proteins that successfully bind to target molecules, with potential for advancing drug design, disease understanding and more. https://buff.ly/3TjIvw7 #bio #biology #synbio
-
This is a sample research design to illustrate the power of the LoopAI Pipeline. In the personalized medicine, multi-omics data (genomics, transcriptomics, proteomics, and metabolomics) reveal how drug combinations affect biological pathways. However, analyzing this complex data to identify key drug interactions and predict outcomes is challenging. The LoopAI Pipeline makes this possible, enabling advancements in personalized treatment strategies.
Automated Analysis of Drug Interaction Effects in Multi-Omics Data for Personalized Medicine with LoopAI App
https://executiveai.uk
-
A new #AI system has been developed to design novel, high-strength protein binders, offering expedited opportunities to understand biological processes, enhance drug discovery, and enable the development of biosensors. #ArtificialIntelligence #HealthResearch #Biology
AlphaProteo generates novel proteins for biology and health research
deepmind.google
-
AlphaFold3 = Protein structure prediction by taking in amino acid sequence. AlphaProteo = Design Protein Binders, by taking in target molecule as an in put. AlphaProteo uses AlphaFold3. #Biology #AI https://lnkd.in/ggMv8HaW
AlphaProteo generates novel proteins for biology and health research
deepmind.google
-
🌟 Excited to Announce the Publication of My Paper! 🌟 I am thrilled to share my latest research: "Machine Learning Models for Predicting Monoclonal Antibody Biophysical Properties from Molecular Dynamics Simulations and Deep Learning-Based Surface Descriptors", now published in Molecular Pharmaceutics. In this study, we introduce AbDev, a biophysical property predictive tool designed to help researchers predict 12 commonly used biophysical properties of monoclonal antibodies (mAbs) directly from their sequence. I am immensely grateful for the support and guidance of my advisor Pin-Kuang Lai, and other co-authors - Lateefat Kalejaye, for their amazing contribution towards this achievement. 💡 AbDev Web App: Explore it here 👉 https://lnkd.in/gHsFfdbv This tool combines molecular dynamics simulations and deep learning-based surface descriptors, aiming to streamline the development of therapeutic antibodies and enhance our understanding of their biophysical properties. Read the full paper here: https://lnkd.in/g7-kh7sa The codes and parameters are equally freely available on Github: https://lnkd.in/ga7acznD #MachineLearning #Biophysics #MonoclonalAntibodies #DeepLearning #AbDev #Research
Machine Learning Models for Predicting Monoclonal Antibody Biophysical Properties from Molecular Dynamics Simulations and Deep Learning-Based Surface Descriptors
pubs.acs.org
-
💥💥💥 Lupex Biotechnologies Unveils Exciting New Development! 💥💥💥 ✨ We are thrilled to unveil our new range of PCR products, meticulously designed to revolutionize efficiency in genome research! 💎 🌏 Explore our latest offerings: 1. Taq DNA Polymerase – ensuring high performance for sensitive clinical applications, including precise DNA amplification and accurate detection in diagnostic assays. 2. Hot Start DNA Polymerase - Enables Hot Start amplification to minimize non-specific binding and improve accuracy, making it perfect for precious samples. 3. Bst Polymerase - features efficient strand displacement and isothermal amplification, making it ideal for LAMP and high GC content regions. 4. EVA Green qPCR Master Mix - Provides exceptional gene amplification, supports high-throughput applications, and enables detailed analysis through high-resolution melt curve profiling. 5. PCR Master Mix - PCR master mix enables efficient DNA amplification, supports long target amplification, detects low copy numbers, and enhances PCR efficiency. 6. Pfu DNA Polymerase - High-fidelity PCR using a low-error rate, thermostable DNA polymerase ideal for applications needing precise base insertion accuracy. 7. KOD DNA Polymerase – is a high-fidelity, high-processivity enzyme, ideal for PCR applications requiring precise and efficient DNA amplification up to 19Kb. 8. phi29 DNA Polymerase - Ideal for multiple displacement amplification with high fidelity, extreme processivity for large fragment generation, and isothermal amplification. 9. KlenTaq1 Polymerase - Perform PCR directly from whole blood samples, eliminating the need for costly DNA extraction and purification, and amplifying from limited quantities of DNA. 10. High fidelity T4 DNA ligase - engineered variant of T4 DNA Ligase with improved thermostability, Capable of joining blunt end and cohesive end termini, as well as repairing single stranded nicks in duplex DNA. 🏆 What makes our new PCR Products stand out? ✅ • Enhanced performance in efficiency, specificity, and sensitivity. • Designed for high-throughput quantitative and qualitative PCR applications. • Amplifies DNA templates across a broad spectrum of sample types. #LupexBiotechnologies #MolecularBiology #PCR #qPCR #Biotechnology #GeneticResearch #Genomics #LifeSciences #DNAAmplification #LAMP #ResearchAndDevelopment #ScientificInnovation #NewProducts #enzymes #bestenzymes #genomicresearch
-
-
Research Scientist - Crop Transformation at MyFloraDNA - Plant Genetic Engineering/Plant TissueCulture/Genomics/Gene Editing/Agrigenomics
Making NEW "Alphabets of Life" in Lab - DNA becomes TNA 🧬In a pioneer-shifting advance, scientists from the University of Cologne (UoC) have developed artificial nucleotides, the building blocks of DNA, with several additional properties in the laboratory, which could be used as artificial nucleic acids for therapeutic applications. 🧬The developed "threofuranosyl nucleic acid" (TNA) is more stable than the naturally occurring nucleic acids DNA and RNA, which brings many advantages for future therapeutic use. 🧬The 5-carbon sugar deoxyribose, which forms the backbone of DNA, was replaced by a 4-carbon sugar, and the number of nucleobases was increased from four to six. By exchanging the sugar, the TNA is not recognized by the cell's own degradation enzymes. 🧬 This overcomes a challenge in nucleic acid-based therapeutics, as synthetically produced RNA that is introduced into a cell is rapidly degraded and loses its effect. 🧬TNAs could also be used for the targeted transport of drugs to specific organs in the body (targeted drug delivery) as well as in diagnostics; they could also be useful for the recognition of viral proteins or biomarkers. Publication: https://lnkd.in/gMAxWFkm https://lnkd.in/gK9SZRn9
Researchers develop artificial building blocks of life
phys.org
-
Exciting news! AlphaFold 3 by Google DeepMind and Isomorphic Labs is revolutionizing molecular biology and drug discovery with groundbreaking accuracy and accessibility. 🔬 Predicts structures and interactions of proteins, DNA, RNA, and ligands 💊 Enhances drug design and discovery processes 🌐 Free and easy-to-use AlphaFold Server for researchers worldwide 🤝 Collaborations with pharmaceutical companies for real-world applications #AI #Biotech #Innovation 🧬 Predicts molecular structures with 50% more accuracy ⏩ Accelerates research in drug discovery and disease understanding 🔍 Provides detailed insights into biological processes 🌍 Offers free access to the AlphaFold Server for transformative research 🎯 Models complex interactions, including protein-ligand and protein-antibody binding 🧪 Assists in developing new treatments by predicting how molecules interact in the human body 🧬 Supports research in neglected diseases and food security through global partnerships 📚 Expands education with a free online course in collaboration with EMBL-EBI Access it now for groundbreaking scientific insights!
AlphaFold 3 predicts the structure and interactions of all of life’s molecules
blog.google
-
AI engineer. Full-stack software engineer. Physician. Interventional cardiologist. #ai #computervision #llm #genai
Google DeepMind announced AlphaFold 3, a new iteration of the AlphaFold AI model aiming at predicting predicts protein foldings and their interactions with DNA, RNA, and other small molecules https://lnkd.in/ey3XQTMX AlphaFold 3 is reported to surpass existing methods by at least 50% It is promised to lead to transformative advancements in pharmaceutical research, materials science, agriculture, genomics In the meantime, like with previous versions of AlphaFold, the release will likely lead to bench tests aiming to see whether the predictions lead to real chemical effects: https://lnkd.in/eu9PuSYb #deepmind #alphafold3 #ai #drugdiscovery
AlphaFold 3 predicts the structure and interactions of all of life’s molecules
blog.google
-
𝗣𝗵𝗮𝗴𝗲 𝗗𝗶𝘀𝗽𝗹𝗮𝘆 𝗨𝗻𝗹𝗲𝗮𝘀𝗵𝗲𝗱: 𝗔𝗜 𝗮𝗻𝗱 𝗡𝗚𝗦 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗶𝗻𝗴 𝗗𝗿𝘂𝗴 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆 𝗮𝗻𝗱 𝗧𝗵𝗲𝗿𝗮𝗽𝗲𝘂𝘁𝗶𝗰𝘀 Phage display is a powerful technique that displays peptides or antibodies on the surface of bacteriophages, enabling the screening of large libraries for molecules with desired properties. The 2018 Nobel Prize in Chemistry recognized George P. Smith and Sir Gregory P. Winter for their pioneering work on this technology. 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗮𝗻𝗱 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝘀 Phage display has been widely used in epitope mapping, drug target identification, and therapeutic development. It is increasingly integrated with next-generation sequencing (NGS) and artificial intelligence (AI) to enhance peptide and antibody discovery. 𝗡𝗼𝘁𝗮𝗯𝗹𝗲 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 Studies using NGS platforms, such as Roche 454 and Ion Torrent, have led to the identification of thousands of peptide targets. Researchers have also applied AI to predict peptide binders, such as PD-L1, improving phage display accuracy and efficiency. 𝗠𝗲𝘁𝗵𝗼𝗱𝗼𝗹𝗼𝗴𝗶𝗰𝗮𝗹 𝗜𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁𝘀 Recent reviews and studies focus on enhancing phage display techniques, such as bioconjugation, non-canonical amino acids, and ELISA signal optimization. Additionally, phage display-derived antibodies have shown efficacy against various RNA viruses, including SARS-CoV-2. 𝗖𝗼𝗻𝗰𝗹𝘂𝘀𝗶𝗼𝗻 Phage display continues to evolve with NGS and AI integration, contributing significantly to biomarker discovery, vaccine design, and drug development. Summarized by Samuel Ndegwa. 𝗥𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲 Huang, J., Yoichi Takakusagi, & Ru, B. (2022). Editorial: Phage display: Technique and applications. Frontiers in Microbiology, 13. https://lnkd.in/dbdST6n4 #PhageResearch #PhageTherapy #OneHealth #PhageDisplay #AntimicrobialResistance
-
