Motivation : In model organisms such as yeast, large database of protein-protein interactions has become an extremely important resource for the study of protein function, evolution, and gene regulatory dynamics. In this paper we demonstrate that by integrating protein interaction data with mRNA expression data, it is possible to identify active functional modules, i.e., gene groups that show significant changes in expression under specific conditions such as nitrogen starvation and other stress conditions. Compared to S.cerevisiae, protein interaction data of S. pombe and newly sequenced H. polymorpha are remarkably small or none. To overcome this problem we proposed pipeline to identify functional modules in active functional modules from predicting protein interaction by integrating mRNA expression data.
Method: We proposed method to find active functional modules (active sub-networks) under certain conditions - especially nutritional restriction such as glucose addtion and nitrogen starvation using PPI data and mRNA expression data. Firstly based on comparative genomics approach, we derived protein-protein interaction map of H. polymorpha and S.Pombe from protein-protein interaction map of other source organism (S. cerevisiae). To infer protein-protein interaction map from source organism, we adopted interacting domain profile pair (IDPP) method  along with naive sequence comparison. Secondly based on the idea that genes with correlated expression changes over many conditions or time courses are likely to be related to similar functions or cellular processes, we proposed combined molecular network model from protein-protein interaction map and mRNA expression association rule map. In the final step, using this combined model we applied simulated annealing method to detect active sub-network that is highly related to certain functional pathway under given condition based on the p-value scoring function .