A plethora of methods and tools have been developed to identify genes differentially expressed between two compared conditions, such as "diseased versus healthy". For many scientists, however, the type of result that has been obtained is not easily interpretable. For example, when comparing a batch of diseased tissue samples to healthy controls, many researchers will interpret the resulting gene list as being "disease-specific". This is actually not the case, since one has only compared samples for one disease and on a single tissue type. The genes obtained are therefore merely "related to a certain degree" to this disease in the tested tissue. To answer the question of "which genes are disease-specific", one needs to include data from many other diseases. This presentation will illustrate this concept using gene expression data from several studies on Crohn's disease and from hundreds of other diseases. The difference between disease-SPECIFIC and disease-RELATED genes will be explained and how such genes can be discovered using a high quality, deeply curated compendium of expression data.