Figure 4.

Algorithmic scheme of PWEA. In step 1, two different colors (yellow and orange) in the signature vector indicate two phenotypes (for example, normal and cancer). Blue rectangles in the gene list vector indicate genes in a particular pathway P_k. For a pathway k, the expression profiles are categorized into two groups: P_k (blue) and its complement, 'Not P_k' (cyan). In step 2 the TIF scores for genes in P_k are calculated. In step 3, TIF scores of the genes in 'Not P_k' set is computed. In step 4, the maximum deviation (MD) between two cumulative distribution functions is computed. After calculating MD for each of n iterations of phenotype shuffling, the fraction of occurrences of shuffled MDs ≥ the original MD is the P-value of P_k. In step 5, after all pathways have been tested, FDR is used to correct for multiple testing. In step 6, results and a KEGG markup language topology file for visualization in visANT [68] are the final output. CDF, cumulative distribution function.