Software helps advance predictive modelling for drug-metabolising enzymes
Optibrium and Lhasa, developers of software and artificial intelligence (AI) solutions for drug discovery and development, have announced the publication of a peer-reviewed study in the Journal of Medicinal Chemistry.
In the paper, ‘Predicting Regioselectivity of AO, CYP, FMO and UGT Metabolism Using Quantum Mechanical Simulations and Machine Learning’, the team combined existing experimental results, quantum mechanics and machine learning to build predictive models for drug metabolism.1 The research will underpin the development of new capabilities for StarDrop, enabling users to determine the metabolic fate of drug candidates and further streamline the preclinical drug discovery process.
Unexpected metabolism can cause the failure of many late-stage drug candidates or even the withdrawal of approved drugs. It is therefore essential to predict metabolism for potential drug candidates. Current predictive models of metabolism usually target the human Cytochrome P450 (CYP) enzyme family due to its well-characterised role in the metabolism of drug-like compounds. However, there is an increasing need to predict metabolism for other enzymes, such as human Aldehyde Oxidates (AOs), Flavin-containing Monooxygenases (FMOs), and Uridine 5’-diphosphoglucuronosyltransferases (UGTs).
The study demonstrates novel predictive models for AO, FMO, and UGT metabolism, and extends the existing model for CYP metabolism to preclinical species. Expanding the portfolio of predictive models beyond CYPs will enable drug discovery scientists to determine a compound’s metabolic fate more accurately, helping to design better drugs and identify toxicity earlier in the project. In silico modelling for CYP in preclinical species can also reduce animal testing in toxicology studies, making trials quicker, less expensive, and more ethical.
Dr Mario Öeren, Principal Scientist at Optibrium, said: “We are delighted to see our research published in The Journal of Medicinal Chemistry. Combining quantum mechanical simulations and machine learning has allowed us to successfully expand predictive models of metabolism to new enzymes — a unique undertaking which addresses some of the key preclinical challenges of today. We are confident that this research’s demonstrated ability to predict metabolism across a broad range of different metabolic enzymes will provide an invaluable resource for scientists approaching drug discovery.”
Dr Matthew Segall, CEO of Optibrium, commented: “A huge congratulations to our team on this achievement. Here at Optibrium, we are always looking to innovate, and we pride ourselves on the scientific rigour behind our portfolio. The research will deliver powerful new capabilities to our StarDrop platform, strengthening our mission to push the boundaries of what’s possible within the computer-aided drug discovery space.”