Modelling β−Trefoil Proteins Using an Object-Oriented Database

This article describes on-going work using an object-oriented database, P/FDM, to build 3D homology models of β-trefoil proteins such as Interleukin-1 (IL-1), by exploiting existing structural data from known structures of the same fold. The model-building procedure is based on the powerful querying capabilities of P/FDM that allows data access, navigation, and computation to be mixed freely and easily in order to formulate complex structural queries. The database can be queried using a Java-based graphical user interface. Using the P/FDM β−trefoil database, the above approach allows to “cut-and-paste” fragments from existing β−trefoil structures to build the model. The structural part of the database contains objects for the description of proteins, as well as β-trefoil structure-specific entities. Data are organised according to a global β-trefoil structural alignment based on superposition of the 18 core residues of the fold, identified from previous structural studies. The database is currently populated with 46 trefoil domains and further structures are being added in order to make available a richer collection of “spare parts” for modelling. Future enhancements such as improved side chain placement and fragment clustering are also described.