Current Research

My current research at Wageningen University & Research (WUR) encompasses several cutting-edge themes in the field of chemometrics and biostatistics. Each theme addresses unique challenges and contributes to the advancement of statistical methodologies for complex data interpretation.

1. Development and Validation of Statistical Models for Predicting Complex Quantitative Traits

A primary focus of my research is developing and validating statistical models to predict complex quantitative traits, particularly through QSAR (Quantitative Structure-Activity Relationship) modeling. This involves creating models that predict the activity of molecules based on their chemical structure. At WUR, in collaboration with the Food Chemistry department, we have developed a robust QSAR methodology. This methodology includes nested cross-validation to ensure model reliability and genetic algorithm-based variable selection to manage the thousands of potential molecular predictors. This approach not only enhances predictive accuracy but also provides insights into the underlying biochemical interactions.

2. Simplivariate Methods

In the realm of high-dimensional data, distinguishing between informative and non-informative variation is crucial. Traditional methods often lump all variation into either systematic or noise categories, which can obscure significant findings. Simplivariate methods, developed from first principles, explicitly formulate the problem by separating data into informative and non-informative parts. This enhances the analysis of metabolomics and other omics data, leading to more interpretable models and uncovering biologically relevant information that might be missed by conventional methods. These methods address the critical challenge of data interpretability in complex biological datasets.

3. Omics-related Research

Continuation of metabolomics research with applications in tomatoes, cocoa, coffee, rice, and human breast milk, addressing statistical challenges such as time dependency and batch correction.