Fig. 1

Overview of the SUBATOMIC workflow. Input: SUBATOMIC takes as input a multi-edge network consisting of directed and/or undirected interactions of different interaction types. Supporting information might contain additional input files such as GO terms, gene annotations, and a list of subgraph definitions that can be used by ISMAGS to screen specifically for these subgraphs (Methods). SUBATOMIC: (1) The multi-edge network is then decomposed into composite subgraphs using ISMAGS for co-pointing (COP), co-regulated (COR), circular (CIR), feed forward loop (FFL), feedback undirected loop (FBU), feedback 2 undirected loop (FB2U), complex (COM) and 2 feedback (2FB) subgraphs. (2) Based on the subgraphs, SCHype generates topological modules for each subgraph type as well as for all subgraph types together (ALL). (3) Modules are connected to each other as well as to regulators to produce an integrated view (further called superview). (4) Modules are functionally annotated with GOATOOLS. (5) We generated files that can be imported into Cytoscape for module visualization. Networks analyzed in this way are considered static if they do not incorporate any condition specific information. Downstream Analysis: Here modules are integrated with condition-specific expression data. Several scores reflect the condition-specific activity of modules: the expression dynamicity score (ECD), the average Pearson correlation of expression values in a module (nPCC) and the module activity score