How To Use PAGI

The section introduces our pathway analysis based on global influence (PAGI) method for identifying canonical biological pathways associated with different biological states. PAGI used a network-based approach to find the latent dysregulated pathways by considering the global influence from both the internal effect of pathways and crosstalk between pathways. Firstly, we constructed a global gene-gene network based on the relationships of genes extracted from each pathway in KEGG database and the overlapped genes between pathways. The global gene-gene network data is stored in the environmental variable (netWorkdata). The expression profiles data with normal and disease samples were mapped to the global network. Then we defined a global influence factor (GIF) to distinguish the non-equivalence of gene influenced by both internal effect of pathways and crosstalk between pathways in the global network. The random walk with restart (RWR) algorithm was used to evaluate the GIF score by integrating the global network topology and the correlation of gene with phenotype (see the section 2.1). We used the function CalGIF to calculate the GIF scores. Finally, we used cumulative distribution functions (CDFs) to prioritize the dysregulated pathways ((see the section 2.2)). We used the function PAGI.Main to prioritize the pathways.