The application of a fully integrated and automated virtual screening method for identifying potential and novel inhibitors of bovine lmwPTP is described. The protocol makes extensive use of our recently introduced LINGO tools, which allow the extraction of the implicit chemical information present in SMILES representations. Nine out of 34 compounds selected from a database of almost 500,000 commercially available compounds were experimentally confirmed to be competitive inhibitors of lmwPTP, two of them showing K(i) values around 10microM. The best inhibitors previously described had K(i) values higher than 1mM. These results constitute an experimental validation of our virtual screening algorithm and provide a basis for the optimization of pharmacologically interesting lmwPTP inhibitors.