Database Entity Relationship Diagram. The top row of each table is the primary key, and all of the relationships are one-to-many. The three tables at the top of the figure, "protein_family", "dyndom_run", and "domain" existed prior to this analysis. Listed in each table are the attribute names together with data type. The names are self-explanatory in most cases. In the "contact-residue" table, the attribute "type" categorises the interaction type (e.g. a hydrogen bond or just a contact), its location (e.g. in an extended bending region), and whether the hydrogen bond is with the side chain or main chain. In the "ligand_analysed" table and the "ligand_conf" table the "contact_type" attribute indicates whether the ligand is spanning or non-spanning. See main text for further explanation.