Modeling & Simulation

Cyprotex has been performing innovative mathematical modelling for more than 20 years with a specific focus on ADME, pharmacokinetics (PK), pharmacodynamics (PD), pharmacology, toxicology and systems pharmacology.

Key developments during this period include:

• The early production of generic physiologically based pharmacokinetic (PBPK) models for compound screening during drug discovery:
  • Application of a generic physiologically based pharmacokinetic model to the estimation of xenobiotic levels in rat plasma
  • Application of a generic physiologically based pharmacokinetic model to the estimation of xenobiotic levels in human plasma
• The development of a world-leading machine learning system for creating quantitative structure activity/property relationship (QSAR/QSPR) and other statistical models, for use in a wide variety of settings and drug discovery applications.
• Development of models for predicting pharmacokinetics and pharmacodynamics (PK/PD) in the hollow fibre infusion model, a flexible in vitro system for simulating in vivo pharmacokinetic profiles.

We specialize in the use of PBPK modeling for prediction of pharmacokinetics in lead identification and early lead optimization. In addition to bespoke PBPK modelling for addressing specific requirements, we offer an innovative screening service to provide reliable PK prediction from early ADME data, enabling compound selection on predicted human pharmacokinetics.

We have many years’ expertise in the application of machine learning to the prediction of ADME, PK and toxicity in drug discovery. Models developed using machine learning are integral components of many of our PK prediction services. These include models for the prediction of properties and activities from compound structure alone (i.e. QSAR and QSPR) for virtual screening, and others integrating structural properties and in vitro data for prediction of complex in vivo properties. We also offer a service for the development of bespoke models using clients’ proprietary data.

Machine learning is primarily performed by a proprietary system developed in house. The system utilises repeated grid-search cross validation to generate robust models with low variance. Our paper describing the core features of this system has received more than 800 online citations.

Our team has extensive experience in PK/PD modelling, particularly across a wide variety of target organisms in the area of anti-infectives. We have experience in modeling in a range of in vivo and in vitro systems, our expertise including:

• Compartmental modelling
• Non-compartmental modelling
• PK driver determination
• Modelling of vascular and extra-vascular PK/PD
• Empirical and systems pharmacology models
• Prodrugs/active metabolites

We have particular and extensive expertise modelling the hollow fibre infusion system for single drugs and for multi-drug combinations.

Systems biology and systems pharmacology are powerful suites of methods for integrating data from multiple in vitro, ex vivo and in vivo sources – whether ADME/PK, toxicity and/or efficacy. This allows the production of mathematical models for predicting the behavior of systems in living organisms, and their responses to external stimuli, such as pharmaceuticals or other xenobiotics. Properly developed systems models can generate valuable additional information from data, enabling improved decision making, cost reduction and reduction in animal usage. Cyprotex has extensive expertise in systems biology and pharmacology, including applications in areas such as:

• Oncology
• Anti-microbials
• Endocrinology
• Energy metabolism