AI and advanced modelling expertise - e.g. DISEASE-03

TU/e research group Computational Biology brings complementary expertise in data-driven modeling, mechanistic simulation, AI-based prediction and virtual human twins. This provides a strong foundation for the development of AI-based risk prediction models, supporting prevention, diagnosis, and management of post-infection long-term conditions.

Predictive modeling and machine learning

We have developed and validated clinical prediction models using machine learning on large-scale clinical datasets (e.g., Mestrom et al 2021, PubMed ID 33621892). This experience ensures robust handling of heterogeneous patient data, feature engineering, and external validation, which is critical for building patient-specific digital twins that integrate real-world clinical trajectories.

Mechanistic modeling and simulation

We have designed multi-organ and multi-scale models of human (patho)pathophysiology, including the inflammatory response to infection (Relouw et al 2024, PubMed ID 39638779), incorporating physiological parameters and patient-specific characteristics. This mechanistic expertise allows integration of biophysical models with data-driven components, supporting hybrid digital twins that combine interpretability with predictive power.

Deep learning for dynamic risk assessment

Our recent work on irregular time-series modeling (GRU-D, Transformer-based architectures) demonstrates state-of-the-art performance for risk prediction in perioperative settings. These methods can also be developed for long-term post-infection conditions, enabling real-time risk stratification and adaptive treatment strategies within a digital twin framework.