Harnessing AI for Enhanced DMPK Modeling
Recorded: 27 May 2025
This webinar will address the different approaches to DMPK modeling, comparing data-driven methods that look for trends in existing data with physics-based models that use fundamental equations. It will also cover the limitations in data availability, such as heterogeneity and bias, and the critical importance of understanding chemical space and domain applicability for accurate predictions. DMPK models will be discussed, including examples of good data management practices. We will also dive into different data splitting techniques and how they impact model performance. Finally, we will explore cutting-edge AI applications in DMPK.
Agenda:
- Understanding the Foundations and Limitations of DMPK Modeling
- Data Curation and AI Model Evaluation in DMPK
- Future Directions: Mechanisms, Interpretability, and Physics in DMPK
Moderated by:
Nicolas Duchemin - Business Development Manager at Pharmaron
Speakers:
Bart Lenselink - Director, Lead Cheminformatics and Data Science Technology at Structure Therapeutics
Bart is a passionate Dutch computational drug discovery expert and currently serves as Director, Lead Cheminformatics and Data Science Technology at Structure Therapeutics.
With a PhD from Leiden University, Bart has notably focused on optimizing G Protein-Coupled Receptor (GPCR) ligand, which was a collaboration with Schrödinger. He continued this work during a postdoc in a joint project between Leiden and Janssen Pharmaceutica. After this, he worked for almost seven years at Galapagos, starting as a scientist and quickly advancing to his final role as team lead of computational technologies. His research has resulted in the publication of over 40 articles in peer-reviewed international journals.
Daniel Price - Vice President, Computational Chemistry and Structural Biology at Nimbus Therapeutics
Dr. Daniel Price is Vice President of Computational Chemistry & Structural Biology at Nimbus Therapeutics, where he leads a team of internal and external scientists focused on delivering breakthrough medicines through structure-based design, leveraging both physics-based and knowledge-based predictive modeling. Before joining Nimbus, he spent 16 years at GlaxoSmithKline, where he led a team of computational chemists and data scientists across various areas, including structure- and ligand-based drug design, high-content screening analytics, predictive modeling, and cross-functional research informatics. Dr. Price received his undergraduate degree in chemical engineering from the University of Colorado at Boulder, followed by his Ph.D. in Molecular Biophysics & Biochemistry from Yale University with Prof. Bill Jorgensen. He completed an NIH postdoctoral fellowship with Prof. Charlie Brooks, III, at The Scripps Research Institute before joining GSK.
John Maclean - Senior Principal Scientist, Computational Chemistry and Informatics at Pharmaron
Dr. Maclean joined Pharmaron in 2017 and has over 25 years of industrial experience, spanning the pharmaceutical, biotechnology, and contract research organization (CRO) sectors. He has extensive expertise in structure-based drug design, initially as a crystallographer, and later as a computational chemist. He supported over 50 projects, from hit ID through lead ID and lead optimization, and has increased the use and impact of structure-based drug design in all his roles. John has expertise in multiple disease areas, including central nervous system (CNS), oncology, immunology, and anti-infectives, and possesses an excellent understanding of the drug discovery process. He has helped teams overcome challenges in selectivity, ADME/PK, and toxicity, and progress seven compounds into clinical development. John’s experience spans the full range of CADD techniques, from ligand-based drug design to structure-based drug design. It includes the design and implementation of informatics infrastructure to support drug discovery projects. He is the author of more than 30 published papers and is named as an inventor on over 35 patent applications.