Figure 2

Examples of the relationship between rmsd and seq.id. For three cases (PDB codes: 1GOZ – S. aureus enterotoxin, 1LAT – R. norvegicus DNA binding domain of the glucocorticoid receptor, 1WRK – H. sapiens N-terminal domain of cardiac troponin C) we show how the quality of the largest cleft in the experimental structure decreases with seq.id. The atoms of the largest cleft in the experimental structures are shown in orange (structures in the left), while the same atoms are shown in red in the various homology models. Cleft quality becomes very poor at seq.id. below 30%, although even at higher seq.id. level some shape details are clearly lost. The distinct templates, as well as the target protein, are identified by their CATH [50] domain identifier, which comprises the four letters of the PDB code, a chain symbol, plus two digits indicating the domain.