Section III works on prediction models and focuses on algorithmic systems biology, using bioinformatics methods and machine learning concepts (PREDICTION). The aim is to better understand and predict the changes and biological effects of biofunctional food effector systems. Their findings could contribute to the development of personalized nutrition concepts in the future.

Research Group: Network Regulation & Modeling / Machine Learning

The research group Network Regulation & Modeling / Machine Learning (focus: network analysis and machine learning) investigates and applies methods from bioinformatics, systems biology, machine learning, and artificial intelligence (AI) to understand interactions between food and consumers at the molecular level. It models biological networks and their regulatory mechanisms using causal system models and validates them through in silico simulations and *omics perturbation experiments using high-throughput sequencing and proteomics. The aim is to build databases (data from Sections I–III) that link foods via molecular components to biochemical networks and cellular phenotypes.

Research Group: Molecular Modeling

The Molecular Modeling research group (focus: modeling and simulation of effector-receptor interactions) assesses, compares, and optimizes new computer-aided tools to investigate the structure and dynamics of molecular target structures and develops methods for efficient and rapid virtual screening. The main research interest lies in the development of predictive models that can be used for the targeted search for new bioactive compounds to drive future food innovations.

Research Group: Integrative Food Systems Analysis

In the Integrated Food Systems Analysis research group (focus: integrative multiomics analyses), complex data sets from innovative high-throughput technologies from sections I and II (e.g., mass spectrometric metabolome, proteome, transcriptome, and genome data) are analyzed in an integrative manner using machine learning methods in combination with network enrichment tools or by applying bi-clustering algorithms in order to predict new effector systems.