Biomemory’s Post

Unlocking the Future of Data Storage with DNA Technology Our lead biochemist, Dr. Corentin R. F., had the pleasure of speaking to the final-year students at the École Supérieure des Biotechnologies de Strasbourg (ESBS) about DNA data storage, a groundbreaking solution to the challenges of long-term data storage. Key takeaways from the session: • The current crisis in data centers and its environmental impact. • Opportunities offered by molecular storage: unmatched stability, exceptional density, and low energy requirements. • The immense potential of DNA to revolutionize how we store and preserve information. Highlights: • Engaging discussions with curious and passionate students. • Exploring the intersection of biotechnology and IT to address real-world challenges. • Sharing insights on a promising technology that could shape the future. A big thank you to the students and faculty at ESBS for this inspiring opportunity! These exchanges reaffirm the importance of innovation and education in driving impactful change. #DNADatastorage #Biotechnology #Innovation #Sustainability #FutureTech

The increasing accessibility of foundation models is transforming AI development for fields like chemistry and drug development! Foundation models are pretrained on massive datasets and are broadly applicable to a range of research challenges. AI teams fine-tune these models using small, labeled dataset to quickly develop high-quality, task-specific models. Recursion’s latest publication introduces their new chemical property foundation model, MolE. MolE’s two-step pre-training process uses molecular graphs to build a versatile model. First, MolE was trained with a self-supervised approach on 842 million molecular graphs to learn general chemical properties. Next, it was further refined with supervised training on 456K molecules with labeled properties. MolE’s adaptability was then benchmarked on 22 ADMET benchmark datasets from the Therapeutics Data Commons (TDC). TDC datasets vary from a few hundred to several thousand molecules. MolE showcases its capability to fine-tune on minimal data while achieving robust performance. Check out the open access article in Nature Portfolio Communications (https://lnkd.in/eH-8qFFY) for a deeper dive into MolE’s development, or explore the code for your own fine tuning experiments (https://lnkd.in/eNBYkqFp)

🎉 We're continuing our celebration of National Nano Day (10/9) this week! Two big problems we face are microbial antibiotic resistance and the buildup of plastic waste. Even worse - the former compounds the latter! We need a new approach to studying bacteria and finding treatments that minimizes waste. Dr. Aeron Tynes Hammack is developing genomic sequencing pipelines that can test individual molecules one at a time. This innovative system will provide insights into the behavior of single cells while using significantly less plastic than current testing methods. In the next episode of Foundry Fireside, join Dr. Hammack as he discusses his work with a talented high school intern from Berkeley Lab's Experiences in Research summer program (run by Berkeley Lab K-12 Programs) 🌟 Watch the full video here: https://lnkd.in/giY725QQ Want more Foundry Fireside? You can watch the entire new season here: https://lnkd.in/gTCHg6da #NationalNanoDay #Nanoscience #STEM #Innovation #Nanotechnology #AntibioticResistance #Plastic #MolecularBiology

Excited to share our last paper published in Chemical Science (Royal Society of Chemistry) which has been selected as part of the 2024 Chemical Science HOT Article Collection and made a journal cover: 𝐖𝐚𝐭𝐞𝐫-𝐆𝐥𝐲𝐜𝐚𝐧 𝐈𝐧𝐭𝐞𝐫𝐚𝐜𝐭𝐢𝐨𝐧𝐬 𝐃𝐫𝐢𝐯𝐞 𝐭𝐡𝐞 𝐒𝐀𝐑𝐒-𝐂𝐨𝐕-2 𝐒𝐩𝐢𝐤𝐞 𝐃𝐲𝐧𝐚𝐦𝐢𝐜𝐬: 𝐈𝐧𝐬𝐢𝐠𝐡𝐭𝐬 𝐢𝐧𝐭𝐨 𝐆𝐥𝐲𝐜𝐚𝐧-𝐆𝐚𝐭𝐞 𝐂𝐨𝐧𝐭𝐫𝐨𝐥 𝐚𝐧𝐝 𝐂𝐚𝐦𝐨𝐮𝐟𝐥𝐚𝐠𝐞 𝐌𝐞𝐜𝐡𝐚𝐧𝐢𝐬𝐦. Check it out here (Open Access): https://lnkd.in/d8YzzBEk To elucidate the molecular mechanisms of SARS-CoV-2 spike dynamics, we performed μs-long all atom AMOEBA high-resolution polarizable adaptive sampling molecular dynamics simulations, zooming into the interaction interface in its both open and closed states. Our findings revealed a sophisticated protein-solvent glycan polarization network essential for stabilizing the open state. Additionally, we demonstrated that the glycan shield consistently preserves viral camoufl age across both conformations, effectively protecting the virus from immune detection. These insights not only deepen our understanding of viral behavior but also pave the way for further exploration of viral evasion strategies. Amazing work by Margaret Blazhynska and kudos to Louis Lagardère and Olivier ADJOUA. Another great collaboration with Chengwen Liu and Pengyu Ren. This research was funded by European Research Council (ERC) (project EMC2). It also benefited from the support of Qubit Pharmaceuticals and GENCI. Large scale #GPU-accelerated computations were performed thanks to a corporate sponsorship from Amazon Web Services (AWS) through the Fondation Sorbonne Université. Special thanks to Gilles TOURPE Brian Skjerven Nicolas Malaval (AWS). #SARSCov2 #covid19 #HPC #supercomputing #drugdesign Sorbonne Université CNRS The University of Texas at Austin

Check out Recursion's latest paper that just got published in Nature Communications! 🎉 We introduce MolE, a new foundation model for chemistry. Paper (open access): https://rdcu.be/dZWa9 Blog: https://lnkd.in/ebGYAcqr Code: https://lnkd.in/eBXU9ETT MolE was trained on a massive set of ~842 million molecules in a self-supervised manner. This means MolE doesn't rely on experimental data to "learn chemistry". But the innovation doesn't stop there. Once pre-trained, MolE showcases its versatility as it can be fine-tuned to a variety of downstream tasks. The adaptability of MolE results in superior prediction capabilities, making it an extremely valuable tool for drug discovery projects. We hope this cutting-edge tool enhances research and discovery processes in the field of chemistry. Finally, I want to thank my coauthors Christos A. Nicolaou and Berton Earnshaw for their support and mentorship. #Chemistry #MolE #Research #Innovation #AI4Science #AI4Chemistry #Cheminformatics

R&D has evolved from solo efforts to dynamic, collaborative teams driving knowledge production. Organizations are now investing in technological infrastructure to support cross-disciplinary teams, speeding up experimentation and discovery. In this #webinar Brian K. Buntz, Editor-in-Chief of R&D World will interview Matt Armbrust, a Research Associate in Protein Engineering for Sarafan ChEM-H | Stanford University, and Jenny Hu of Labstep, a STARLIMS company, to discuss: 🔬 The current state of scientific experimentation from practitioners, and how R&D has changed throughout the years. 🔬 How organizations are breaking down barriers between R&D labs and scientific disciplines. 🔬 Strategies for collaborating more effectively within and across labs and scientific disciplines. Register now: https://lnkd.in/g8wW2Kxd #laboratories #research #researchanddevelopment #crossdiscipline

On training a chemistry foundation model. Innovation in the service of #drugdiscovery.

This year’s #NobelPrize laureates in chemistry have revealed proteins’ secrets through computing and artificial intelligence. Chemists have long dreamed of fully understanding and mastering the chemical tools of life – proteins. This dream is now within reach. 2024 chemistry laureates Demis Hassabis and John M. Jumper have successfully utilised artificial intelligence to predict the structure of almost all known proteins. This year’s chemistry laureate David Baker has learned how to master life’s building blocks and create entirely new proteins. The potential of their discoveries is enormous. The ability to create proteins that are loaded with new functions is just as astounding. This can lead to new nanomaterials, targeted pharmaceuticals, more rapid development of vaccines, minimal sensors and a greener chemical industry – to name just a few applications that are for the greatest benefit of humankind.

In case you could not make it to the Science Laboratory Technology Virtual Open House, you can catch the speakers in the video. Use the chapters to go directly to the speaker who interests you. https://lnkd.in/gXpDmh57

🌍 Where would we be without science and engineering? Yesterday, I had the pleasure of an interesting conversation at the Technical University of Denmark (DTU), which reminded me of an old LinkedIn post of mine: “In my next life, I want to be an engineer.” The conversation reminded me of the synergy between biological scientists and engineers: 🔬 Biologists use cutting-edge technologies developed by engineers to explore and understand the world around us. ⚙️ Engineers, in turn, apply their expertise to craft new solutions that help address unanswered biological questions. The result? A continuous cycle where new discoveries prompt new technologies, fueling even more discoveries. This dynamic is thriving in Copenhagen's biotech and pharma ecosystems—and I love being part of it! Just for fun I prompted an AI image generator based on above text (slightly edited). Hightlighting how AI will surely (and hopefully) accelerate drug discovery in the next 10 years. Another P.S. I realize it's a bit old-fashioned to distinguish between these fields, especially with the rise of hybrid disciplines like biomedical engineering, biomaterials science, bioinformatics, and bioprocess engineering. But the core message remains: collaboration drives innovation. Original post: https://lnkd.in/denFBX2f