Mining the Protein Data Bank to Differentiate Error from Structural Variation in Clustered Static Structures: An Examination of HIV Protease