Gene Networks from Microarray Data

Since the development of the microarray technique in 1995, there has been an enormous increase in gene expression data from several organisms. Based on the view of gene systems as a logical network of nodes that influence each other's expression levels, scientists dream of being able to reconstruct the precise gene interaction network from the expression data obtained with this large scale arraying technique. Computer science shows that inference of a logical regulatory network is possible solely from sets of expression data, and mathematicians are working on the question how much data is at least necessary for reverse engineering.

Meanwhile, experimental biologists are experiencing problems in the field. The number of experiments that are necessary before attempting network reconstruction is a lot more than is generally possible in ``wet'' laboratories, so data compression algorithms are applied to reduce the number of nodes considered. This is however an extremely coarse representation of the intricate interconnections that exist between single genes. The resulting network of only a handful of nodes is therefore usually only sufficient to describe the experiments performed, while any possible predicting properties are absent.

In this literature thesis, I attempt to give an update on the state of the art in computerised network reconstruction techniques, and explicitly relate this to actual biological gene networks. I will go into the model formalisms used to describe genetic networks, and explain their specific advantages and disadvantages. Also, a separate chapter will be dedicated to several experimental results obtained in the research of genetic networks, and finally, a short discussion and some hypothesising is added.