Figure 1

The data model of the Genopolis Database. The data model of the Genopolis Database. We represent data by means of a business objects layer. The representation is consistent with the MIAME recommendations and is similar to MAGE-OM. Respect to this two main simplifications have been introduced. First, we ignore chips manufacturing, since we make use of only one standard platform already described by Affymetrix. Second, we model the experiment design as a tree structure. Although this is not general (for instance, the same stimulus could be applied to two samples) it is a good compromise between flexibility and easy of use. We use controlled vocabulary and protocol classes to manage non free text fields.