Assessment of Ligand and Protein Criteria

A number of circumstances led to the removal of putative ligands (i) non-protein macromolecule, for example, DNA or RNA (ii) commonly occurring non-ligands, for example, glycerol and tartaric acid, that were automatically identified and excluded on the basis of their PDB residue ID (GOL and TAR, respectively) and (iii) molecules 

This triage led to the selection of around 26 600 ligand-protein combinations from the 70 700 X-ray crystal structures available as of December 2010. This stage was the most time consuming, as automated selection could not be applied with complete confidence. Inevitably the amount of human input required also reduced the objectivity of the selection, so improvements in this part of the work are definitely possible. 

---