iOmicsPASS: network-based integration of multiomics data for predictive subnetwork discovery

a iOmicsPASS workflow. iOmicsPASS takes multiomics data, biological network data, and sample meta information as input. The omics data sets are integrated via interaction scores for all interactions in the network. Subnetwork discovery module discovers the subnetwork signatures distinguishing phenotypic groups, and pathway enrichment module reports associated biological processes. The software also produces a set of text files containing the details of the selected subnetworks and the materials for visualization of networks in the Cytoscape software. b Each omics data is first standardized into Z-scores and converted to interaction scores over the network. Two TFs (gene 1 and gene 3) and their common target gene (gene 2) are shown as an example. Interaction scores are computed for the PPI between protein 1 and protein 3 and the transcription factor regulation between the two TF proteins and mRNA molecule of their target gene. c The resulting interaction scores are used as an input to select the predictive edges for phenotypic groups using the modified nearest shrunken centroid algorith