An alternative to enrichment methods are module-based approaches, which examine sets defined by other biological characteristics for meaningful pathways contained therein. For example, one study used hierarchical clustering to form modules of co-expressed genes across multiple inflammatory diseases; subsequent analysis of these modules suggested a role for interferon-inducible signaling in tuberculosis [56]. Gene modules can also be defined through protein interaction networks, as in a study that associated genetic variants in glutamate pathways to brain glutamate concentration in multiple sclerosis [57]. Importantly, recent studies are combining enrichment and module-based methods to point to broader findings. For example, network analysis of enriched pathways revealed major roles for antigen presentation and interferon signaling in rheumatoid arthritis [58].
