Several of our projects require simultaneous analysis of gene expression both at the mRNA and protein levels. Samples (in the data shown spleen leukocytes) are subjected to extraction by a trireagent (Trizol™) to yield purified protein and RNA from each sample. Proteins are subjected to proteome analysis using 2D-gel electrophoresis. RNA is subjected to transcriptome analysis on Affymetrix™ expression arrays. Collection and normalization of expression data from the two platforms follow standard procedures. The Affymetrix expression array data were filtered to exclude genes for which none of the samples has a expression (or marginal) call by the MAS software.

Analysis was conducted separately on transcriptome and proteome data, and only at the last step of data mining were the results “merged”. The first step in analysis of the data was a statistical analysis (ANOVA) to define genes that show difference in expression levels between predefined groups in the dataset. This was followed by hierarchical clustering of the resulting list of genes to define clusters of genes whose expression patterns follow specific patterns relative to already defined sample groups or clusters that may define new subgroups within the already known groups.

Gene lists were created from each cluster of interest and subjected to data mining by two main methods. First gene ontologies (GO) were evaluated to see if specific biological processes may be present with excessive frequency. In the second step the Ingenuity™ Pathway server is used to create networks from each list based on actively acquired information from the literature. Particular emphasis were focussed on identifying the most centrally connected genes (nodes) in those molecular networks.

Although we found very little overlap between lists of genes that were differentially expressed at the mRNA and at the protein levels, the central genes in networks created from lists of mRNA and proteins had a remarkable overlap with respect to the central controlling genes found in those networks.