Why model?

Abstract

Next generation sequencing technologies are bringing about a renaissance of mining approaches. A comprehensive picture of the genetic landscape of an individual patient will be useful, for example, to identify groups of patients that do or do not respond to certain therapies. The high expectations may however not be satisfied if the number of patient groups with similar characteristics is going to be very large. I therefore doubt that mining sequence data will give us an understanding of why and when therapies work. For understanding the mechanisms underlying diseases, an alternative approach is to model small networks in quantitative mechanistic detail, to elucidate the role of gene and proteins in dynamically changing the functioning of cells. Here an obvious critique is that these models consider too few components, compared to what might be relevant for any particular cell function. I show here that mining approaches and dynamical systems theory are two ends of a spectrum of methodologies to choose from. Drawing upon personal experience in numerous interdisciplinary collaborations, I provide guidance on how to model by discussing the question "Why model?"

Keywords: cell biology; mathematical modeling; systems biology; systems medicine.

Figure 1

Answering biological questions through mathematical analyses. The table illustrates a selection of approaches available in systems biology. In practice, the vast majority of questions in experimental biology concern “differences.” Given that experimental observations vary, statistical testing will establish the significance of a difference. The experiments for this type of question are easy to conduct but little more than establishing a difference is possible. At the other end of the spectrum models of dynamical systems allow investigations about causal mechanisms underlying complex interaction networks. These very powerful explanatory models do however require sufficiently rich quantitative time course experiments, which in many cases are time-consuming, expensive and technically more challenging.