Figure 1

Overview of the data analysis pipeline. (1) Protein interactions from public databases are collated to build an integrated human PPI network. (2) Greedy unsupervised clustering is used to identify relevant, possibly overlapping, submodules in the PPI network. (3) Hits from one or several RNAi screens are mapped to these modules and modules are filtered for significant enrichment. (4) Subnetworks are further filtered based on network topology and semantic similarity values. (5) Resulting modules are visualized as subnetworks, color-coded for hits, non-hits, and (6a,b) are then functionally characterized based on GO and Reactome pathway. (6c) Lastly, using gene expression data from different tissues, tissue-specific putative novel host factors are predicted.