Using large-scale perturbations in gene network reconstruction

It was found that a relatively small number of perturbations significantly improve inference accuracy, particularly for low-order inputs of one or two genes. The perturbations themselves should alter the expression level of approximately 50–60% of the genes in the network.Time-series obtained from perturbations are a common form of expression data. This study illustrates how gene networks can be significantly reconstructed from such time-series while requiring only a relatively small number of calibrated perturbations, even for large networks, thus reducing experimental costs.Recent technological advances have led to an explosive growth in high-throughput genomic and proteomic data such as DNA microarrays. The rapid growth in available data has led in turn to a need for novel quantitive methods for analysis. As a consequence of this need, the reconstruction of gene network architectures from DNA microarray expression data has become a major goal in the field of systems biology. An increased understanding of the network architectures and their respective dynamics will enable novel approaches to disease treatments by allowing us, for example, to identify drug targets in silico which manipulate the functional outputs of these networks. This process is expected to lead to novel classes of drug based on a network approach to cellular dynamics.Frequently, the gene expression data itself is derived from perturbation experiments such as stress conditions, temperature shifts, and chemical treatments; for example, the widely used yeast cell-cycle datasets of Cho [1] and Spellman [2]. Although these global perturbations are carried out in order to reveal causality between genes, it is not always clear how experiments should be designed so as to reveal as much causality as possible, while both minimising costly experimentation and remaining computationally tractable.A range of computational and mathematical techniques have been adopted in the effort to find a successful gene ne