Fig. 5

Interactome of altered genes in CCM-deficient models with four validations. Only one gene (node) was matched that was identified as perturbed in four different CCM studies. For datasets, non-human genes were converted to human homologs before processing and all genes capitalized to ensure overlaps were not misidentified due to case changes. Interactomes were constructed using Cytoscape software equipped with STRING application. Interactomes illustrate affected pathways with query protein and ten interactors being incorporated into the interactome with a false discovery rate (FDR) set to 0.4 due to the smaller number of query proteins. No post modifications were made to the layout of the interactome once generated in STRING and exported to Cytoscape. Crystal structures, if available, for identified proteins are shown within the corresponding nodes