Integration of multilevel OMICs data based on the identification of regulatory modules from protein-protein interaction networks

Complex scientific experiments provide researchers with a wealth of data and knowledge from heterogeneous sources. Analyzed in its entirety, OMICs data provide a deep insight into the overall biological processes of organisms. However, the integration of data from different cellular levels (e.g., transcriptomics and proteomics) is challenging. Analyzing lists of differentially abundant molecules from different cellular levels often results in a small overlap, which can be accounted to, e.g., different regulatory mechanisms, different temporal scales as well as inherent properties of the measurement method. Thus, there is a need for approaches that allow efficient integration of OMICs data from different cellular levels. In this study, we make use of transcriptome, proteome and secretome data from the human pathogenic fungus Aspergillus fumigatus challenged with the antifungal drug caspofungin. Caspofungin targets the fungal cell wall leading to a compensatory stress response. We analyze the experimental data based on two different approaches. First, we apply a simple approach based on the comparison of differentially regulated genes and proteins with subsequent pathway analysis. Second, we compare the cellular levels based on the identification of regulatory or functional modules by two module-detecting algorithms from protein-protein interaction networks in conjunction with transcriptomic and proteomic data. Our results show that both approaches associate the fungal caspofungin response with biological pathways like cell wall biosynthesis, fatty acid metabolism as well as carbohydrate metabolism. Compared to results of the simple approach, the use of regulatory modules shows a notably higher agreement between the different cellular levels. The additional structural information of the networks provided by the module-based approach allows for topological analysis as well as the analysis of the temporal evolution of cellular response at a molecular level. However, we also found that quality of the module-based results depends on the comprehensiveness of the underlying protein-protein interaction network itself. Thus, while our results highlight the benefits and potential provided by a module-based analysis of OMICs data from different cellular levels, future studies will have to focus on the expansion of organism specific protein-protein interaction networks.