DIA assay libraries enable holistic functional analyses of co-regulated clusters
Significance testing is a commonly used method for evaluating proteomic data, but often fails to illuminate the context in which these proteins are functioning. The DIA assay library generated in this study contained over 2000 proteins, which permitted non-biased clustering of all corresponding protein abundance patterns associated with salinity acclimation. In our example, cluster analysis identified groups of proteins which were regulated in similar ways to the significant proteins, capturing 88% of the significant proteins in one of two clusters and expanding the number of proteins to be evaluated by six times while still focusing on proteins responding similarly to a salinity challenge. The value of this expansion was particularly evident in the STRING network analysis, as an analysis of only 42 significant proteins did not return any protein-protein networks with more than one edge. The expanded list returned a complex network indicating the connections between significant proteins with non-significant intermediaries, and the effects of significant regulation on connected proteins despite these effects not reaching the level of statistical significance. Additionally, STRING and KEGG enrichment only returned one protein domain and one KEGG pathway which were highly enriched for significantly up-regulated proteins. Thus, our study is in agreement with previous reports pointing out that systems scale “omics” studies are rendered more powerful when a variety of threads of evidence, including cluster analysis, are used to expand on significance testing results (Gehlenborg et al., 2010).