Weekly TechBio News: ADMET Prediction
Startups for AI-Guided Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) Prediction
Startups and novel Tools for AI-Guided Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) Prediction
Startups 🗂️: Aganitha AI, Deep Mirror, DevsHealth, DeepCure, SimulationPlus, HeartBeatBio, Elix, Genesis Therapeutics, Standigm, Quris, Iktos, GreenStone Biosciences and many more
Tools 🗂️: DeepDelta, SAFIRE, ADMET-AI and many more
“Life is like a prism. What you see depends on how you turn the glass.”
By Jonathan Kellerman
🔮 Aganitha AI Inc
Aganitha (India, 2017) is a new generation in silico company that integrates high-throughput sciences with DL based generative models to solve complex drug discovery and development challenges. Their approach focuses on deep sciences and deep tech collaboration and includes:
Multidisciplinary teams with expertise in AI, DL, high-throughput sciences and computational chemistry and biology to accelerate innovation,
Blending in silico techniques with wet lab approaches, using physics-based chemistry and first principles-based systems biology,
Applying advancement in modalities, including gene and cell therapies, mRNA/RNA, advanced antibody engineering, and ASO, to develop more precise and effective therapeutics.
Utilizing methods and techniques to handle data deluge (e.g., Omics) and data gaps (e.g., antibody data for AI/ML) to discover novel therapeutics.
Offering the following
Solutions: Antibody Engineering, GWAS Pipeline, mRNA Platform, Omics Pipelining Reaction Modeling, SMOL Drug Design, ADMET predictions, Virtual Screening and Crystal Structure Prediction. And
Services: Large Language Model Services, Computational Biology Computational Chemistry, Technology and Cloud Services, Proteomics and RNA sequencing.
On May 29, 2024, the Centre for Cellular & Molecular Biology (CCMB) under the Council of Scientific & Industrial Research (CSIR) and Aganitha signed an umbrella memorandum of understanding (MoU) to apply Generative AI solutions for designing novel therapeutics and research tools addressing needs in multiple disease areas. Initial areas of collaboration include target analysis, small molecule design, antibody and nanobody engineering for addressing Malaria, Tuberculosis (TB) and neurological disorders (CSIR-CCMB and Aganitha sign a framework agreement to apply Generative AI for therapeutic design and research in multiple disease areas).
Aganitha was founded by Vikram Duvvoori and Prasad Chodavarapu.
🔮 Deep Mirror Ltd
DeepMirror (founded by a team of researchers at the University of Cambridge in 2019) makes AI enabled drug design as simple as spreadsheets, and empowers every biopharma team to find better drugs faster. Its user-friendly interface enables medicinal chemists of all levels to deploy this powerful approach in a fraction of the time. The ability to apply DeepMirror’s platform to any desired endpoint, whether it be potency, selectivity, or even ADME properties, empowers its users to make more informed decisions and to do so faster.
When they put the DeepMirror Engine to the test against the top algorithms on the TDC leaderboard (accessible open-source algorithms and an alternative commercial solution) the DeepMirror Engine showed a remarkable ability to provide top accuracy while automatically and dynamically adapting to very different ADMET property prediction tasks, offering the best overall predictive performance for this set of ADMET property predictions (DeepMirror Engine put to the test in the Therapeutics Data Commons benchmark). The performance was assessed on test sets provided by the TDC for each dataset, ensuring that the data used for testing was not involved in training. They used a total of 9 benchmarking datasets for which prediction performance metrics are available from individuals or groups that submit their entries to the leaderboard (https://tdcommons.ai/benchmark/ADMET_group/overview/).
Andrea Dimitracopoulos, Co-founder & Product Lead at DeepMirror, just announced that during their latest collaboration with Medicines for Malaria Venture (MMV) (a Swiss-based non-profit at the forefront of antimalarial drug research), where they joined forces to improve one of their antimalarial compound series, DeepMirror enabled MMV to find a new compound (in just one hour) with high predicted antimalarial activity and lower predicted CYP450 inhibition than their lead molecule (Using DeepMirror to accelerate the fight against malaria). MMV's lead molecule in their aryl piperazine series faced a significant challenge: it could cause adverse interactions with other drugs by inhibiting the CYP450 enzyme.
Accordingly, they downloaded the aryl piperazine dataset from the Malaria Libre data repository and uploaded the molecules to DeepMirror. Then they generated >1,000 novel compounds with their generative AI capabilities, with similar antimalarial activity and lower CYP450 activity. In just one hour, DeepMirror identified a novel compound (DM1133) with predicted CYP450 activity above 10 µM, a potential improvement of ~10x over the existing top candidate. Importantly, these predictions were experimentally validated by MMV's testing entities. This success showcases how DeepMirror's unique combination of generative and predictive capabilities can accelerate the journey from idea to clinic.
🔮 DevsHealth SL
DevsHealth is a Spanish DeepTech company using AI, Real-World Data (RWD) and molecular modeling to improve anti-infectious treatments. Their AI-based technology helps to optimize a rational design of new drugs, anticipate potential side effects and its behavior in our body by predicting ADME properties. By integrating, standardizing and curating several public-source databases at DevsHealth they are able to manage around 2.5M gene expression experiments, almost a million bioactive compounds and their biological activities and thousands of proteins and structures.
The foundation for DevsHealth OS is Data Galaxy based on curated data from public sources as well as their own generated data, providing a vast and invaluable collection of real-world information. This extensive database empowers DevsHealth OS to continually enhance and develop groundbreaking drug candidates for infectious diseases.
On July 4, 2024, IGTP and DevsHealth joined forces to develop new antibiotics against hospital-acquired infections by multi-resistant Staphylococcus aureus. IGTP, Germans Trias i Pujol Research Institute, is a public research center in biomedical sciences. During this collaboration IGTP will provide DevsHealth with access to almost 400 bacterial strains, including their whole genome and antibiotic resistance profile. This will offer DevsHealth the opportunity to work with a massive data analysis to find new mechanisms of action that can avoid or overcome resistance by using its DeepTech platform, which integrates AI, real-world data and molecular modeling, to run an in silico drug discovery process in only a few days and generate novel molecules with optimal pharmacological and safety profiles, as well as predict their efficacy and toxicity.
On September 24, 2024, DevsHealth announced a strategic partnership with Aragen for the chemical synthesis of its projects. Aragen is a leading company in the synthesis of chemical compounds, headquartered in India and with offices in Europe and North America. This strategic partnership establishes Aragen as DevsHealth’s primary synthesis provider and partner for all chemical synthesis requirements.
🔮 DeepCure Ltd
DeepCure (Boston, Massachusetts, US, 2018) is a startup that uses DL to discover small molecule therapeutics. Their proprietary PocketExpander™ uses AI and physics-based methods to identify physical and chemical features on the protein surface that can interact with small molecules. The technology includes quantum mechanical simulations to calculate binding energies and molecular dynamics simulations of protein folding.
Their molecular generation tool, MolGen™, generates custom libraries of compounds that are designed to selectively interact with features identified by PocketExpander™. The process leverages state-of-the-art deep reinforcement learning (RL) to build readily synthesizable molecules using millions of available building blocks and more than 200 robust chemical reactions. MolGen™ also ensures that each molecule has the desired ADME-Tox profile and target candidate profile (TCP). MolGen™ generates novel compounds that are readily synthesizable and can selectively interact with the features identified by PocketExpander™ as well as meet ADMET and TCP requirements.
Their DiscoveryHub™ platform was created to enable sharing of data and molecule designs between structural biologists, computational chemists, AI engineers, medicinal chemists, biologists and project managers. The platform includes tools to design, visualize and evaluate compound structures as well as plan the optimal synthetic route for each molecule selected for synthesis.
On April 11, 2024, DeepCure Closed $24M Funding Round Led by IAG Capital Partners. On May 28, 2024, DeepCure announced that its Inspired Chemistry™ platform had achieved a breakthrough in chemical synthesis synthesizing: nirmatrelvir and 56 analogs in parallel using a robot-driven workflow. On August 19, 2024, DeepCure linked with Leeds Institute to bring AI-based rheumatoid arthritis (RA) drug (DC-9476) to clinic. The two entities will conduct a study using blood samples and joint biopsies from various patient subgroups, focusing on those resistant to current treatments and they will then leverage advanced translational tools to analyze gene expression patterns and cytokine levels in patient cells and tissues. On September 17, 2024, DeepCure Presented In Vivo Efficacy and Safety Data of BRD4 (BD2) Inhibitor DC-9476 for Macrophage Activation Syndrome at EMBO Conference: DC-9476 blocked macrophage activation syndrome (MAS) disease development in the CPG-induced animal model of MAS. On October 03, 2024, DeepCure Presented In Vivo Data Showing Its Selective BRD4 (BD2) Inhibitor DC-9476 is Superior to Etanercept in Rheumatoid Arthritis: BRD4 (BD2) inhibitor DC-9476 is superior to anti-TNF-α treatment in the collagen antibody-induced arthritis (CAIA) rheumatoid arthritis (RA) mouse model.
DeepDelta: predicting ADMET improvements of molecular derivatives with deep learning
DeepDelta provides an accurate approach to predict molecular property differences by directly training on molecular pairs and their property differences to further support fidelity and transparency in molecular optimization for drug development and the chemical sciences.