Figure 3.

Representative examples from the module content similarity analysis. The significance of the similarity in module content between ReMoDiscovery seed modules and GRAM [9] and SAMBA [3] output is shown at different parameter settings. The color bar on the right indicates the normalized Jaccard similarity score, that is, the number of standard deviations from the mean of the distribution of Jaccard similarity scores on randomized module partitioning. (a) Regulator content similarity between ReMoDiscovery and GRAM, with varying GRAM module p value cutoff and ReMoDiscovery Chip-chip threshold. (b) Gene content similarity between ReMoDiscovery and GRAM, with varying GRAM core profile p value cutoff and ReMoDiscovery correlation threshold. (c) Gene content similarity between ReMoDiscovery and SAMBA, with varying SAMBA overlap prior factor and ReMoDiscovery correlation threshold.