MetaNetwork Enhances Biological Insights from Quantitative Proteomics Differences by Combining Clustering and Enrichment Analyses

Interpreting proteomics data remains challenging due to the large number of proteins that are quantified by modern mass spectrometry methods. Weighted gene correlation network analysis (WGCNA) can identify groups of biologically related proteins using only protein intensity values by constructing protein correlation networks. However, WGCNA is not widespread in proteomic analyses due to challenges in implementing workflows. To facilitate the adoption of WGCNA by the proteomics field, we created MetaNetwork, an open-source, R-based application to perform sophisticated WGCNA workflows with no coding skill requirements for the end user. We demonstrate MetaNetwork's utility by employing it to identify groups of proteins associated with prostate cancer from a proteomic analysis of tumor and adjacent normal tissue samples. We found a decrease in cytoskeleton-related protein expression, a known hallmark of prostate tumors. We further identified changes in module eigenproteins indicative of dysregulation in protein translation and trafficking pathways. These results demonstrate the value of using MetaNetwork to improve the biological interpretation of quantitative proteomics experiments with 15 or more samples.

Automated summary

Interpreting proteomics data is challenging due to the large number of proteins quantified by mass spectrometry methods. Weighted gene correlation network analysis (WGCNA) can identify biologically related proteins by constructing protein correlation networks. To facilitate the adoption of WGCNA in proteomics, we developed MetaNetwork, an open-source application that allows users to perform sophisticated WGCNA workflows without coding skills. Using MetaNetwork, we successfully identified groups of proteins associated with prostate cancer and found dysregulation in protein expression and pathways.