PDBLIG: classification of small molecular protein binding in the Protein Data Bank

It is known that proteins can adopt different folds while sharing similar features for recognition of similar substrates or ligands, for example, in the binding sites of enzyme cofactors such as ATP. On the other hand, proteins that have highly flexible binding sites or belong to large and diverse protein families can bind structurally dissimilar ligands, as, for example, in the case of the matrix metalloprotease family. We have developed a database, PDBLIG, that classifies protein domains and ligands. The information stored includes each protein's function, domain class(es), which ligand(s) it binds, and so on. The database can provide valuable knowledge for drug discovery, supporting the answering of questions such as whether the same drug molecule can bind different target protein families and whether these families are related functionally or structurally, which ligand classes (such as metabolites or organic molecules) bind to a particular protein family and whether the ligands are druglike, and which target families bind a wide variety of ligands and whether different ligands are associated with different subfamilies.