gene and protein interaction networks can be used to infer ontologies whose coverage and power are comparable to those manually curated by the GO consortium (10). We have also built GLAD4U (11), a tool that derives and prioritizes gene lists from PubMed literature for a user provided search term, which opens the possibility to create gene sets related to concepts such as diseases and drugs. Despite their obvious importance, disease and drug-related gene sets are scarce in the curated databases. For example, the average number of genes for each disease in the OMIM (Online Mendelian Inheritance in Man) database (12) is 2.7, making it unsuitable for enrichment analysis. Thus, computationally derived gene sets may serve as an important data source to enhance functional enrichment analysis. The new version of WebGestalt integrates data from centrally and publicly curated databases as well as computational analyses, leading to a total of 78 612 functional categories (Figure 1).
