Figure 3

Visualization of edge weight and the in- and out-degree of genes and TFs in ensembles of sputum networks. (A) Edge weights for every possible transcription factor to gene interaction, where each row represents an edge, and each column represents one of the networks produced in the jack-knifing approach. Rows are ordered based on the t-statistic comparing the edge weight values and each row is Z-score normalized for visualization purposes only. (B) The in-degree, defined as the sum of all incoming edge weights, for each gene in the PANDA network reconstruction. Genes (rows) are ordered based on the t-statistic comparing the gene in-degree distributions in the two ensembles of networks (columns). Again, rows are Z-score normalized only for visualization purposes. (C) The twenty-five most differentially-targeted genes, identified as having the most significant difference in in-degree in the male compared to the female ensemble of networks. Both the significance of the differential-targeting and the level of differential-expression is shown. (D) The out-degree, defined as the sum of all outgoing edges, for each transcription factor in the reconstructed networks. Rows represent transcription factors and are again ordered based on the t-statistic comparing the distribution of in-degree values of the transcription factor in the two ensembles of networks. (E) The ten most differentially-targeting transcription factors, and their level of differential-expression. The majority of the differentially-targeted genes and differential-targeting transcription factors are not differentially-expressed.