Gene set enrichment analysis (GSEA)

GSEA ranks all of the genes in the dataset based on differential expression. To test the gene set significance, an enrichment score is defined as the maximum distance from the middle of the ranked list. Thus, the enrichment score indicates whether the genes contained in a gene set are clustered towards the beginning or the end of the ranked list. Both self-contained and competitive hypothesis tests can be conducted with GSEA by altering how randomization is completed for hypothesis testing. For a self-contained hypothesis the phenotype labels are permuted while the genes are permuted for a competitive hypothesis. A total of 1000 permutations are performed to estimate the empirical p-values for the gene sets. Details of GSEA can be found in Subramanian et al. [49].