The superfamily of enzymes known as the cytochromes P450 (P450s) comprises a wide-ranging class of proteins with diverse functions. They are known, amongst other things, to be involved in the hormonal regulation of metabolism and reproduction, as well as having a major clinical significance through their association with diseases such as cancer, diabetes and hepatitis. Knowledge of the three-dimensional (3D) structure of a protein gives insight into its function. The 3D structures of P450s are therefore of considerable scientific interest. A number of high-resolution structures of P450s have been determined by X-ray crystallography and studies of these structures have provided valuable insights into the mechanism of these enzymes. Only one of these structures is mammalian and as yet there is no structural information on human P450s in the public domain. Until such a structure is solved it is necessary to employ alternative methods to gain structural insight into how human P450s perform their biological function. Here we report on the use of comparative modelling to predict the structure of human P450s based on knowledge of their amino acid sequences plus the 3D structures of other (not human) P450s. As an illustrative example of these techniques we have modelled the structure of P450 2C5 using five bacterial P450 structures as templates. We examine the importance of selecting suitable templates. obtaining a good amino acid sequence alignment, and evaluating the models generated. To improve the quality of the models an iterative cycle of sequence alignment, model building, and model evaluation is employed. The result is a model with excellent stereochemistry, good amino acid side chain environment properties, and a Calpha. trace similar to the crystal structure. (C) 2002 Elsevier Science B.V. All rights reserved.