Maikel Verouden

Introduction

Maikel Verouden studied Environmental Analytical Chemistry at the Hogeschool van Arnhem en Nijmegen, where he received his BSc. degree in 1998. During the final year of his studies he did his internship and research project at Tauw Milieulaboratorium in Deventer. His internship concerned the development of a screening method for the determination of the biochemical oxygen demand (BOD) in waste water and his research project the development of a microwave assisted digestion of waste water for the determination of quicksilver.

After having obtained his bachelor degree he continued working at Tauw Milieulaboratorium till 2000 participating in several research and development projects. In 2000 he decided to continue his studies to obtain his master degree in chemistry. Intrigued by his courses in statistics and chemometrics during his bachelor studies he decided to do his studies at the Universiteit van Amsterdam, because of the presence of the workgroup Process Analysis and Chemometrics (PAC) headed by prof.dr. A.K. Smilde. At the end of the master studies Maikel did his research project in the Biosystems Data Analysis workgroup (formerly PAC), where he studied the validity of the Hammett equation for the Heck coupling reaction of n-butylacrylate with p-substituted iodobenzenes. He also wrote his literature thesis in this group about microarrays. In 2005 he received his MSc. degree.

In november 2005 Maikel started his PhD-project on "Grey and dynamic models for the analysis of metabolomics data of Lactococcus Lactis".

PhD Project

An important aspect of functional genomics is data analysis. There is, however, a serious problem in the analysis of -omics data, namely the risk of spurious results, also called overfitting or chance correlation. The enormous number of variables measured on only a few objects makes this problem severe and hampers obtaining high-quality information from the data. There are several ways to combat this problem, but in this PhD project the use of additional biological information will be explored.

One of the ways to do this is by using grey models. There are several different types of models: On the one hand there exist black-box models (e.g. principal component analysis or neural nets) that use only the data. On the other hand, there are first-principles or mechanistic models (white models) that can be use to model (parts) of a biological system. "Grey models" is used as a generic term for hybrid models – in-between black and white – to model the data. Such models combine the advantages of both black models (good fit) and white models (generating fundamental insight).

Apart from solving the overfitting problem, grey models also provide a vehicle for moving between low-level modelling and high-level modelling. Moreover, they can be used for pathway building and modelling dynamic phenomena. Different versions of grey models will be developed and tested on data that becomes available from the other partners in this BioRange project (NBIC).

The model organism chosen in the BioRange project is Lactococcus Lactis, as a prime example of an industrially important microorganism.