Analysis of a typical gene microarray dataset has customarily involved determining a list of genes that were differentially expressed between several conditions, such as treated vs. control, or tumor vs. normal, or between tumor subtypes. There is then the challenge of drawing useful biological inferences and productive lines of follow-on research from the gene list. One approach for a systematic exploration of the data has been to statistically evaluate whether certain pathways or Gene Ontology categories are significantly overrepresented in the list of differentially expressed genes. These types of methods, however, leave unused the potentially valuable information available in the expression data for all the genes in the platform. More recently, led by the Gene Set Enrichment Analysis (GSEA) method of Subramanian et al. 2005 and Mootha et al. 2003 developed at the Whitehead and Broad Institutes, techniques have been formulated that directly test for overall differential expression of specific sets of genes as a group such as pathway gene sets, e.g., from BioCarta and KEGG, or sets of genes co-located within cytobands or having common transcription factor motifs. These techniques evaluate the expression levels of the genes in a gene set as an ensemble, and utilize all the available gene expression data. In this seminar we describe four leading methods for this type of analysis (GSEA, PAGE, SigPathway and GSA), and present the results of applying them to two semi-simulated data sets in which the expression levels of a designated group of gene sets were varied in controlled ways.