Decoding Cellular Dynamics in Epidermal Growth Factor Signaling Using a New Pathway-Based Integration Approach for Proteomics and Transcriptomics Data

Identification of dynamic signaling mechanisms on different cellular layers is now facilitated as the increased usage of various high-throughput techniques goes along with decreasing costs for individual experiments. A lot of these signaling mechanisms are known to be coordinated by their dynamics, turning time-course data sets into valuable information sources for inference of regulatory mechanisms. However, the combined analysis of parallel time-course measurements from different high-throughput platforms still constitutes a major challenge requiring sophisticated bioinformatic tools in order to ease biological interpretation. We developed a new pathway-based integration approach for the analysis of coupled omics time-series data, which we implemented in the R package pwOmics. Unlike many other approaches, our approach acknowledges the role of the different cellular layers of measurement and infers consensus profiles and time profile clusters for further biological interpretation. We investigated a time-course data set on epidermal growth factor stimulation of human mammary epithelial cells generated on the two layers of RNA and proteins. The data was analyzed using our new approach with a focus on feedback signaling and pathway crosstalk. We could confirm known regulatory patterns relevant in the physiological cellular response to epidermal growth factor stimulation as well as identify interesting new interactions in this signaling context, such as the regulatory influence of the connective tissue growth factor on transferrin receptor or the influence of growth arrest and DNA-damage-inducible alpha on the connective tissue growth factor. Thus, we show that integrated cross-platform analysis provides a deeper understanding of regulatory signaling mechanisms. Combined with time-course information it enables the characterization of dynamic signaling processes and leads to the identification of important regulatory interactions which might be dysregulated in disease with adverse effects.