Although this analysis was by necessity performed on the same data set from which networks were inferred, the Bayesian gene network inference method used does not utilise information about the effects of the siRNA treatments on individual probe sets. Correlations between these parents and their children are significantly larger then correlations between randomly chosen nodes. These children show a trend to be downregulated by parent knockdown when parent and child correlate positively across the dataset, and to be up-regulated by parent knockdown when parent and child correlate negatively. The regulation of child abundance after parent knockdown was generally small in magnitude, consistent with the expected dilution of the effect of knocking down any single parent by the undiminished effects of the remaining parents that were not knocked down. This Bayesian network method was primarily used in this study to identify co-expressed clusters rather that directional regulation. Future experimental evaluation of directional network predictions will be interesting but this is beyond the scope of this study. Sequencing the genome of humans and rodents has provided an immense set of uncharacterized genes, and within the past decades several genetic approaches have been taken in order to address their function. Embryonic stem cells are pluripotent cells that have served as a powerful tool to study gene functions in vitro and to generate knockout mice via homologous recombination. In order to complement data gained from loss-of-function approaches, in vivo gain-of-function experiments have been carried out by generating mice overexpressing a gene of interest. Gain-offunction mouse models have been mainly generated by pronuclear microinjection and random integration of the transgene into the genome. This quite often results in variable copy numbers, unpredictable expression profiles and sometimes gene silencing effects, therefore requiring extensive characterization of several LY2157299 supply independent transgenic lines. Thus, insertional mutagenesis and the positional influence of endogenous genes and regulatory elements often lead to misinterpretation of the phenotypes observed. Targeting a single-copy transgene to a specific and well-defined locus can minimize these problems and provide a predictable and reproducible expression profile. The Rosa26 locus has been used to drive ubiquitous gene expression from the Rosa26 promoter. This locus offers an open chromatin configuration in all tissues and disruption of the Rosa26 gene produces no overt phenotype, which made it one of the most commonly used genetic loci for targeted transgenesis. MK-1775 in vivo However, targeting transgenes to the endogenous Rosa26 promoter results only in moderate ubiquitous expression and is not suitable for high expression levels.