I have a set of genes following network-based methodologies including STRING and ClueoGO. At first these profiles were extracted from GEO and I conducted meta-analysis method to integrate different independent data sets.
Here, I have 4 gene sets from 4 different phenotypes. The size of gene sets are around 100 per integrated data sets. Further, I gonna find master regulators from defined gene sets. Of course, I know this issue about identifying master regulators or disease-causing genes can be quite tough and hard to unify. I am still new to bioinformatics but I'd tried to apply network inference methods including mutual information and regression with R package. However, I found the performance was not good than I expected. I guess one of the reason may be small sample (~50).
I am looking for simple and proper method to find master regulators. Today I've searched bayesian methods but I did not find the easy to follow R package to identify master regulators and reconstruct network and visualize it. Is there any way to find master regulators not integrate other levels of information but use just gene names.
Of course, based on the specific phenotype such as cancer, the backbone network can be different from each other before importing specific gene sets. I found simple method and cytoscape plugin (KDA key driver analysis: http://sagebase.net/research/tools.php). But this seems to be suitable for yeast data.
What am I supposed to do?