Integrated Mechanistic & Data-driven Modeling Group
Hence, the mechanistic & data-driven modeling group focuses on:
- Design and interpretation of clinical trials
- Inform personalized medicine
- Model-informed drug development
- Integrated Mechanistic modelling
- Pharmacokinetic/Pharmacodynamic (PKPD) modelling
As our understanding of human biology deepens, it becomes clear that health and disease are governed by complex, interacting processes that evolve over time. A central challenge in medicine is therefore not only to explain these processes, but to predict their future development: how a disease will progress, how a treatment will modify that progression, and why patients respond differently to the same intervention.
Our group develops predictive models of disease progression and drug response by integrating mechanistic knowledge of biology with patient-level data. We use mathematical models to represent key biological processes and combine them with data-driven approaches that capture inter-individual variability and longitudinal patient trajectories. By combining these perspectives, we generate quantitative predictions that are directly relevant for drug development and clinical decision-making, including applications in precision dosing and digital twins.
This translational focus connects quantitative models, biomarkers, and clinical data to better understand and predict treatment effects across individuals and over time. It also shapes our teaching and supervision, where quantitative models are introduced as tools for reasoning about biological and clinical data, rather than as purely technical exercises.
The response of biological systems to therapeutic interventions is often complex, involving interactions across multiple pathways and scales. Predicting how drugs act in individuals or populations requires robust, mechanistic models that capture this complexity. Our research focuses on developing quantitative methods that not only simplify and enhance these models but also expand the ways we can study and apply them in practice.
A key aspect of our work is model reduction, where complex systems pharmacology (QSP) models are distilled into more interpretable and computationally efficient forms. This allows researchers and clinicians to better understand treatment outcomes, explore multiple scenarios, and integrate mechanistic knowledge with pharmacokinetic and pharmacodynamic (PKPD) models to guide decision-making.
Equally central is the development of synthetic dataset generation. Synthetic data have emerged as a powerful solution to challenges such as data scarcity, privacy concerns, and the need to train AI models with sufficient statistical power and unbiased coverage of patient variability. In healthcare—and particularly in PKPD modeling—realistic synthetic datasets are still rare, due to limitations of current methodologies in capturing the complex, multimodal structure of patient data. Our work aims to overcome these limitations by generating statistically robust, physiologically plausible datasets that reflect real variability in patient populations. These datasets provide a reproducible and shareable resource, moving away not only from the “one-size-fits-all” paradigm in dosing, but also from the traditional “one dataset—one model” approach. In the future, this framework will enable both models and datasets to be published, supporting transparency, reproducibility, and broader applicability in research and clinical practice.
Through this dual focus on model simplification and advanced synthetic dataset generation, our methods provide rigorous, flexible, and clinically relevant tools. They form a foundation for predictive, reproducible, and AI-enhanced pharmacometrics research that
Patients respond differently to drugs due to biological, genetic, and environmental factors. Understanding these differences is critical to designing effective and safe treatments. Our research combines traditional pharmacometric modeling with artificial intelligence (AI) to:
- Predict early treatment resistance
- Identify disease progression patterns
- Optimize dosing strategies for individual patients (precision medicine)
We apply AI to PKPD modeling by integrating data from clinical trials, biomarkers, and patient monitoring to inform treatment decisions. These models help clinicians tailor therapies, improve patient outcomes, and accelerate drug development. For example, we use AI-informed PKPD models to identify early markers of treatment resistance or to adapt dosing strategies as patient responses evolve over time.
By linking AI with mechanistic pharmacometric models, we enable precision medicine strategies that are not only predictive but actionable. This work provides opportunities for the next generation of researchers and students to engage in a highly interdisciplinary environment, combining mathematics, statistics, pharmacology, biology, and clinical science to tackle real-world challenges in patient care.
PhD Students/Postdocs
We are always interested in supporting motivated students and researchers on their scientific journey. Are you a mathematician interested in learning how to apply mathematics in health science?
Are you a pharmacist eager to learn how to take your programming and statistical understanding to the next level to improve patients’ lives? We are a highly interdisciplinary group and welcome many backgrounds, united in the aim to work toward the future landscape of healthcare.
Please send over a letter of intent outlining your motivation to join, as well as publication list and CV.
Here available, fully funded PhD or postdoc positions will be posted: https://employment.ku.dk
MSc/Exchange/Scholarship students
We welcome inquiries from a broad range of backgrounds, reflecting the highly interdisciplinary nature of our group.
Here current project available can be found: https://drug.ku.dk/education/bachelor-and-master-projects/translational-pharmacology-master-projects/
Group Leader
Jane Knöchel
Assistant Professor
Phone +45 35 32 21 29
jane.knoechel@sund.ku.dk
Publication list on the University of Copenhagen Research Portal