Skip to main content
Fig. 5 | BMC Genomics

Fig. 5

From: ManiNetCluster: a novel manifold learning approach to reveal the functional links between gene networks

Fig. 5

Comparison of ManiNetCluster with other clustering methods. a The adjusted rand index between ManiNetCluter clustering and other methods, as shown, indicates ManiNetCluster is consistent with k-means and WGCNA but less so with expectation maximization and hierarchical clustering. b comparison of 60 cross-condition modules detected by ManiNetCluster as well as 34 light period modules and 30 dark period modules separately detected by WGCNA by constructing a network, consisting all ManiNetCluster and WGCNA modules as nodes. The links between two nodes indicate the genes shared by both modules. Node size indicates the degree of that node. Links with very low weight are omitted. The triad of the network among three different kinds of nodes (i.e., ManiNetCluster module, WGCNA “light-period” module and WGCNA “dark-period” module) indicates the functional linkage type of an ManiNetCluster module. An open triad patterns indicates a functional linkage module. c Subgraph of the network in b demonstrating a functional linkage module (Module 55). The subgraph also identifies a putative functional link between two WGCNA modules, Light-Module 9 and Dark-Module 14

Back to article page