Biological knowledge management and gene network analysis: a heuristic road to System Biology

In order to understand the molecular basis of living cells and organisms, biologists over the past decades have been studying life's core molecular players: the genes. Most genes have a specific function, a role they play in the collective task of developing a cell and supporting all the aspects of keeping it alive. These genes do not perform their function randomly. Instead, after billions of years of evolution, nature's trial-and-error process, they have become parts of an utterly complex and intricate network, an interconnected mesh of genes that comprises signal detection cascades, enzymatic reactions, control mechanisms, etc. Over several past decades, experimental molecular biologists have sought mainly to study these genes via a one-by-one approach. However, with the advent of high-throughput experimental techniques, the number-crunching power of computers, and the realisation that many biological functions are the result of interactions between genes or their proteins, Biology's related field of Systems Biology has emerged. Here, one tries to combine the dispersed information produced by many researchers, in integrated assemblies called gene networks. Our research comprises the development of two new methods for improved information integration in the field of molecular Systems Biology. The first one aims to support an approach to acquire insights in the dynamics of gene networks (the behaviour of gene activities over time), called 'modelling and simulation' of genetic regulatory networks. Our second new method approaches the problem of how to collect and manage the information necessary to compose such genetic networks in the first place, based on scattered information in a dispersed and increasingly fast growing body of publications. These two methods form two separate parts in this thesis (chapters 2-4, and chapters 5-7). Chapter 1, section 1.3 provides an introductory, complete overview of this thesis. It is intended as a light introduction to my doctoral research, presented in an informal and entertaining way, and mainly addressed to my friends and family. It forms an introduction for the laymen to our work and the concepts that are important for this thesis. Chapters 2, 3 and 4 constitute Part 1 of this thesis. Chapter 2 gives a review of the various formalisms for modelling and simulation of gene networks, as a thorough background for our work presented in the following chapter. Chapter 3 describes SIM-plex, our new software tool that forms a bridge between a mathematical gene network modelling formalism, and the biologist, who usually is more an expert in the biology behind the gene network than a mathematician can ever be. It shields off the mathematics in a new way so as to enable biologists to experiment with modelling and simulation themselves. Chapter 4 describes the various applications that SIM-plex was used for. The research described in Part 2 of this thesis, chapters 5, 6 and 7, emerged from our own need for a better management of biological information. We experienced this necessity while we were building a larger genetic network for the Arabidopsis cell cycle, and it forms a general problem in biology. Chapter 5 gives a background of the currently existing methods for harvesting literature information, but comes to the conclusion that no existing automated or manual method displays sufficient potential to capture the largest part of information from literature in a structured way. In chapter 6, we describe our bold proposal of a new method to tackle this problem: MineMap, a community-based manual text-curation initiative. We describe the various aspects required to make such a project possible, based on our own experiences with our prototype application MineMap. This research is organised in a 'heuristic' way, in the sense that we built a first sketch and a working solution that also generated experiences for improvements in a next design. While chapter 6 describes our new ideas and concrete implementations in considerable detail, chapter 7 then illustrates the core concept behind MineMap