RECEPTOR.AI

Congratulations to our CSO Dr. Semen Yesylevskyy and our scientific advisor Prof. Mikhail Bogdanov for getting a grant support for studying molecular mechanism of lipid-induced topological switching of membrane proteins! This ambitious collaborative project between IOCB Prague and UTHealth Houston is funded by The Ministry of Education, Youth and Sports of the Czech Republic (MŠMT). It is widely accepted that membrane proteins always maintain a single orientation and topology in the membrane, but some proteins, such as bacterial sugar transporter protein LacY, reversibly change their orientation in the membrane in response to changes in the lipid environment or charge of their extracellular domain. Such flipping proteins are appealing molecular models of the membrane proteins with incorrect membrane topology, which are known to be associated with multiple human diseases such as Cystic fibrosis, Alzheimer's disease and sickle-cell anemia. In addition, flipping membrane proteins may have broad practical applications in drug discovery, synthetic biology and applied molecular biotechnology. This study will bring us much closer to understanding the universal principles of protein topogenesis in health and disease. #drugdiscovery #biotech #biotechnology

Last week, I had the opportunity to attend the Boltz-1 Seminar hosted by Jeremy Wohlwend and Gabriele Corso from MIT.  It was an inspiring event and a great example of how biomolecular interaction modeling can be made accessible to everyone. Jeremy and Gabriele introduced Boltz-1, an open-source model built by reworking the AlphaFold architecture. By making both the code and model weights available on GitHub, Boltz-1 provides a strong alternative to proprietary tools like AlphaFold3, giving researchers around the world access to advanced computational methods. 𝗞𝗲𝘆 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻𝘀 𝗼𝗳 𝗕𝗼𝗹𝘁𝘇-𝟭: ◾ New MSA Pairing Algorithm: Boltz-1 features an enhanced multiple sequence alignment (MSA) pairing method that accounts for MSA density. This improvement is crucial for accurate protein-protein interaction predictions, leading to more reliable modeling outcomes. ◾ Unified Cropping Algorithm: Combining spatial and contiguous cropping strategies, Boltz-1 optimizes the way complexes are resized for model training. This unified approach ensures consistency and enhances the model's overall performance. ◾ Updated Pocket-Conditioning: By incorporating one-hot vector representations of pocket information, Boltz-1 becomes more robust against incomplete or imprecise pocket definitions. This makes the model versatile and reliable across various research scenarios. 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗮𝗻𝗱 𝗹𝗶𝗺𝗶𝘁𝗮𝘁𝗶𝗼𝗻𝘀: Benchmark results from the CASP15 evaluation demonstrate that Boltz-1 matches the performance of Chai-1, showing its potential as a competitive tool in the field. However, like other models, Boltz-1 has some limitations, such as output hallucinations when dealing with data that differs substantially from its training set. It is unable to account for the dynamic nature of proteins and specific cases such as isoforms and mutants. Additionally, with an inference time of about one minute due to its heavy architecture, Boltz-1 may not be ideal for high-throughput applications where speed is crucial. In contrast, for specialized tasks like molecular docking, Receptor.AI’s ArtiDock enables focused, high-throughput screening capabilities. Due to the nature of its training on the proteome-wide MD simulations dataset and millions of artificial protein-ligand interactions, ArtiDock outperforms more generalized models like AlphaFold3 in scenarios that require the rapid and precise screening of millions of compounds. All in all, I believe Boltz-1 is an important milestone in the direction of the democratization of biomolecular interaction modeling in this industry. Jeremy and Gabriele did an impressive job, and I’m eager to see where their technology will go in the near future. You can read about Boltz-1 in their preprint: https://lnkd.in/eR4GgVfe #boltz1 #mit #molecularmodelling #drugdiscovery #ai #biotech #pharmainnovation #opensource

We are enhancing ArtiDock, our state-of-the-art Molecular Docking Model, even further! ArtiDock is our best-in-class AI docking model. It continues to be a cornerstone of our AI-driven multiplatform ecosystem, driving precision and efficiency in our drug design projects. We’re excited to announce upcoming upgrades for ArtiDock, our advanced molecular docking model: ◾ 𝗘𝘅𝗽𝗮𝗻𝗱𝗶𝗻𝗴 𝗖𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀 𝗳𝗼𝗿 𝗔𝗹𝗹𝗼𝘀𝘁𝗲𝗿𝗶𝗰 𝗦𝗶𝘁𝗲𝘀 𝗮𝗻𝗱 𝗣𝗿𝗼𝘁𝗲𝗶𝗻-𝗣𝗿𝗼𝘁𝗲𝗶𝗻 𝗜𝗻𝘁𝗲𝗿𝗮𝗰𝘁𝗶𝗼𝗻𝘀: We're developing specialized modules optimized for predicting ligand binding modes in allosteric protein pockets and protein-protein interaction sites, further extending ArtiDock's robust capabilities. ◾ 𝗘𝗻𝗵𝗮𝗻𝗰𝗶𝗻𝗴 𝗣𝗲𝗽𝘁𝗶𝗱𝗲 𝗕𝗶𝗻𝗱𝗶𝗻𝗴 𝗠𝗼𝗱𝗲 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻𝘀: Building on our already strong performance, we're introducing new features to further improve ArtiDock's accuracy in predicting peptide ligand binding modes. ◾ 𝗘𝗻𝗵𝗮𝗻𝗰𝗶𝗻𝗴 𝗗𝗮𝘁𝗮 𝗔𝘂𝗴𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻: We are to further refine our data augmentation strategies, particularly in the generation of artificial pockets. We plan to create even more realistic training data and elevate ArtiDock's performance to new heights. We are proud of the continuous advancements in ArtiDock and remain dedicated to enhancing its capabilities to support cutting-edge drug discovery. Stay tuned for these exciting updates as we strive to deliver even greater precision and efficiency in our drug design projects. Discover the previous ArtiDock version: https://lnkd.in/eXrGCyYE #ArtiDock #drugdiscovery #aidocking #biotechnology #ai #artificialintelligence #pharmainnovation #biotech

Can Generative AI replace traditional Molecular Dynamics simulations? Despite significant advancements, why do machine learning models face challenges similar to traditional methods, and how can researchers overcome them? Our latest article discusses this hot topic in drug discovery, exploring successful applications, technological nuances, and our industry insights. Discover how AI is transforming this field, and connect with us to share your thoughts or collaborate! #ai #md #moleculardynamics #artificialintelligence #generativeai #drugdiscovery #biotech #pharmainnovation #ml #machinelearning

Hiring Alert: Lead the Future of Drug Discovery with Us! At Receptor.AI, we're leading the next generation of drug discovery with cutting-edge multiplatform ecosystem. Our expert team is at the forefront of TechBio, and we're seeking exceptional professionals to join us. 𝟭. 𝗦𝗲𝗻𝗶𝗼𝗿 𝗖𝗼𝗺𝗽𝘂𝘁𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗖𝗵𝗲𝗺𝗶𝘀𝘁 𝗘𝘅𝗽𝗲𝗿𝘁𝗶𝘀𝗲: Proven success in leading in silico campaigns, from virtual screening to IND enablement in drug discovery projects, including developing generative AI models for de novo compound design and structure-based ML models for virtual screening. 𝗦𝗸𝗶𝗹𝗹𝘀: Proficiency in building AI-based drug discovery workflows using advanced computational chemistry tools for virtual screening, molecular docking, and pharmacophore modeling. 𝗘𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲: Success in advancing compounds from hit identification to lead optimization in commercial drug discovery projects, with contributions to IND submissions and patent filings. 𝟮. 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗠𝗮𝗻𝗮𝗴𝗲𝗿 – 𝗗𝗿𝘂𝗴 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆 𝗕𝗮𝗰𝗸𝗴𝗿𝗼𝘂𝗻𝗱: Extensive experience managing large-scale AI drug discovery or biotech projects at the preclinical stage. 𝗦𝗸𝗶𝗹𝗹𝘀: Proficient in managing multiple projects in parallel, analyzing experimental data, and providing strategic oversight of AI-driven projects. 𝗘𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲: Leading interdisciplinary teams, especially in rapidly advancing fields like AI drug discovery and biotech; engaging with pharmaceutical companies, regulatory bodies, advisors, and internal stakeholders to communicate progress, challenges, and outcomes. 𝟯. 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗟𝗲𝗮𝗱 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗣𝗿𝗼𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆: Expertise in Python and Django for scalable back-end development, with a strong focus on building and deploying complex machine learning pipelines, including deep learning architectures (e.g., Diffusion Neural Networks, Transformers, GANs). Proven experience with large language models (LLMs), advanced data handling for big data, and high-performance computing (HPC) for large-scale simulations. 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 𝗗𝗲𝘀𝗶𝗴𝗻: Extensive experience designing complex systems, including microservices and distributed architectures with Docker and Kubernetes, and optimizing database structures (SQL, NoSQL, Graph Databases). 𝗗𝗲𝘃𝗢𝗽𝘀 & 𝗖𝗹𝗼𝘂𝗱: Expert in AWS, GCP, or Azure, with proficiency in CI/CD processes for efficient software delivery. 💡 Interested? We look forward to welcoming talented individuals to our technology-driven team! To apply or learn more about these positions, please contact ai@receptor.ai Join us and shape the future of drug discovery! #hiring #drugdiscovery #ai #biotech #innovation #medicinalchemistry

Receptor.AI Opens New Headquarters in Boston, Placing Itself at the Heart of the U.S. Market! As we expand our U.S. presence, Receptor.AI has established new headquarters in Boston, Massachusetts, while maintaining our existing headquarters in London, United Kingdom. This strategic move positions us at the center of one of the world’s leading biotech and pharmaceutical hubs. "Expanding to Boston allows us to collaborate more closely with our innovation-driven U.S. partners and immerse ourselves in a vibrant biotech community," said Alan Nafiiev, CEO of Receptor.AI. In conjunction with Receptor.AI’s growth, we have expanded our team to 35 members, including industry veterans who have successfully brought numerous drugs to market. Their expertise will be invaluable in further strengthening our competitive advantage in the market and aligning our technologies with industry needs. 👉 Read more about our move here: https://lnkd.in/eUEbCfPu We look forward to engaging in cutting-edge R&D collaborations with the world’s most innovative organizations in Boston. #Boston #Cambridge #AI #DrugDiscovery #Biotech #ArtificialIntelligence #Innovation #BiotechHub

Meet our Business Development Executive, Myroslav Uiazdovskyi, at #BIOEurope on November 4-6, in Stockholm, Sweden. As BIO-Europe celebrates 𝟯𝟬 𝘆𝗲𝗮𝗿𝘀 of fostering partnerships in the life sciences industry, we invite you to explore new collaboration opportunities with us. Discover our AI-powered multi-platform ecosystem, comprised of: 𝗜𝗻𝗱𝘂𝗰𝗲𝗱 𝗣𝗿𝗼𝘅𝗶𝗺𝗶𝘁𝘆 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Advancing therapeutic precision through engineered protein-proximity interactions. 𝗗𝗿𝘂𝗴 𝗖𝗼𝗻𝗷𝘂𝗴𝗮𝘁𝗲𝘀 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Crafting stable and selective drug conjugates with cutting-edge design. 𝗖𝗼𝗻𝘃𝗲𝗻𝘁𝗶𝗼𝗻𝗮𝗹 𝗟𝗶𝗴𝗮𝗻𝗱𝘀 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Designing small and medium molecules that modulate activity with exceptional accuracy. We are looking forward to connecting with you onsite and virtually! 🔍 Learn more about BIO-Europe here: https://lnkd.in/dnuNZ2D #AI #DrugDiscovery #Stockholm #Innovation #Biotech 

𝗣𝗮𝗿𝘁𝗻𝗲𝗿𝘀𝗵𝗶𝗽 𝗔𝗹𝗲𝗿𝘁: Receptor.AI and Reaxense Are to Launch AI-Driven Focused Libraries We are happy to announce that RECEPTOR.AI and Reaxense Inc. have partnered to launch innovative AI-driven focused on-demand libraries for drug discovery, targeting an impressive 𝟴,𝟴𝟬𝟭 𝗽𝗿𝗼𝘁𝗲𝗶𝗻𝘀. This collaboration combines Receptor.AI's advanced AI capabilities with Reaxense's expertise in custom synthesis, accelerating the discovery of new therapeutics and offering comprehensive solutions to the pharmaceutical and biotechnology sectors. Leveraging Receptor.AI's 𝗔𝗜-𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗺𝘂𝗹𝘁𝗶-𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺 𝗲𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺, the library design integrates its various components—from protein structural analysis to the multitask model for ADMET prediction. Compounds are meticulously selected from a 𝟲𝟬 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 𝘃𝗶𝗿𝘁𝘂𝗮𝗹 𝗮𝗻𝗱 𝘀𝘁𝗼𝗰𝗸 𝗰𝗵𝗲𝗺𝗶𝗰𝗮𝗹 𝘀𝗽𝗮𝗰𝗲, including those generated using direct 𝗶𝗻-𝗽𝗼𝗰𝗸𝗲𝘁 𝗔𝗜 𝗺𝗼𝗹𝗲𝗰𝘂𝗹𝗮𝗿 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 techniques. Each library features potent modulators supported by 38 ADME-Tox parameters and 32 physicochemical and drug-likeness parameters. All compounds are provided with docking poses, affinity scores, and activity profiles, offering comprehensive insights into their potential. "Leveraging our advanced AI-driven ecosystem alongside Reaxense's synthesis expertise, we are able to develop libraries specifically designed to meet the most challenging therapeutic needs," says Alan Nafiiev, CEO of Receptor.AI. "This partnership offers comprehensive solutions to the pharmaceutical and biotechnology sectors, accelerating the discovery of new therapeutics," states Dr. Yaroslav Bilokin, Associate Director at Reaxense Inc. Together, we're accelerating the journey from molecular design to real-world therapeutics! 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿 𝗼𝘂𝗿 𝗙𝗼𝗰𝘂𝘀𝗲𝗱 𝗟𝗶𝗯𝗿𝗮𝗿𝗶𝗲𝘀: https://lnkd.in/ei9ZHKRB  Read the full Press Release here: https://lnkd.in/eMAfqQ4H  Learn more about Reaxense on their website: https://www.reaxense.com/ #AI #DrugDiscovery #Biotech #Innovation #Partnership

Meet our CEO, Alan Nafiiev, at #BioFuture2024 in New York City! Alan will be at BioFuture from October 28 to 30, connecting with innovators and investors in healthcare. Discover our 𝗔𝗜-𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗺𝘂𝗹𝘁𝗶-𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺 𝗲𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺, comprised of: • 𝗜𝗻𝗱𝘂𝗰𝗲𝗱 𝗣𝗿𝗼𝘅𝗶𝗺𝗶𝘁𝘆 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Advancing therapeutic precision through engineered protein-proximity interactions. • 𝗗𝗿𝘂𝗴 𝗖𝗼𝗻𝗷𝘂𝗴𝗮𝘁𝗲𝘀 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Crafting stable and selective drug conjugates with cutting-edge design. • 𝗖𝗼𝗻𝘃𝗲𝗻𝘁𝗶𝗼𝗻𝗮𝗹 𝗟𝗶𝗴𝗮𝗻𝗱𝘀 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Designing small and medium molecules that modulate activity with exceptional accuracy. Contact Alan to schedule an in-person meeting at the event or set up a virtual session. Learn more about BioFuture 2024 here: https://biofuture.com/ #AI #DrugDiscovery #NewYorkCity #Innovation #Biotech

🏥 2 clinical candidates: Optimizing Liposome and Lipid Nanoparticle Formulations for TLR Agonists 𝗕𝗮𝗰𝗸𝗴𝗿𝗼𝘂𝗻𝗱: Receptor.AI collaborated with a biotechnology company to optimize the Liposome and Lipid Nanoparticle (LNP) formulations of two drug candidates targeting Toll-like receptors. This collaboration facilitated the transition of the candidates to the IND stage by improving drug load and achieving optimal nanoparticle sizes. 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲:   Liposome and lipid nanoparticle drug formulations faced insufficient drug load caused by suboptimal lipid composition. Optimizing lipid composition and size is critical to balance bilayer stability and drug incorporation while ensuring non-immunogenicity and potency. Using computational techniques, Receptor.AI aimed to improve lipid formulation parameters for clinical application and evaluate each drug composition for aggregation in the solvent. 𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵: A total of 24 different lipid and drug compositions were computationally screened with a range of Receptor’s AI-powered Molecular Dynamics (MD) simulation techniques to evaluate their potential for improving drug load and forming stable nanoparticles, including: • Spontaneous drug incorporation into the bilayers • Spontaneous self-assembly of drug-lipid mixtures • Free energy of drug incorporation • Drug aggregation in the solvent 𝗥𝗲𝘀𝘂𝗹𝘁𝘀: Subsequent experimental validation of three recommended lipid compositions confirmed a remarkable 20-fold increase in drug load compared to the initial composition. Simulations revealed that stable nanoparticles of desirable sizes could be formed using a single compound alone or a mixture of both drug candidates at a remarkably high 1:1 lipid-drug molar ratio. These findings guided the ongoing experimental optimization of nanoparticle formulations to fine-tune their size and stability for clinical use. 𝗖𝗼𝗻𝗰𝗹𝘂𝘀𝗶𝗼𝗻: Receptor AI’s team is proud to have tackled this challenge with a mixture of proprietary AI-powered Molecular Dynamics techniques. Notably, working with lipid formulations was relatively new for us, but our refined tech stack allowed us to advance two clinical candidates successfully. #AI #DrugDiscovery #Formulations #Biotech #ArtificialIntelligence #ClinicalTrials