Abstract
Many individual drug development companies and independent parties have attempted the task of creating computational approaches to predict, and thus help explain, the metabolism of drugs. The cytochrome P450 family is responsible for >90% of the metabolism of all drugs, so the computational prediction of metabolism can help to design drugs and indicate possible drug interactions in the early phases of the drug discovery process. The generation of structures produced by the first step during drug metabolism was investigated using experimental data and by identifying common structural trends, using and cross referencing the following properties: molecular mass, Log P, pKa, cytochrome catalyst and the distance from the site of reaction to electronegative points in the structure. The creation of structural templates with the distances to the site of reaction offered guided predictions to the plausibility of a particular reaction occurring. By comparing the number of total possible reactions available to a molecule to the number of reactions which would occur according to template theory, the average reduction in reaction outcome is 36.1%; with structures expecting to have 15 or more reactions having a reaction reduction of 40.9%. These results show that template theory has a positive effect on removing unlikely reactions and thus narrowing the number of predicted structures for first step metabolism and can aid in further clarifying the metabolic path of a drug for the drug design process. The software generated in this work utilized existing and created new specific functions for the template and produced a reasonable platform for the template theory to be hosted and eventually expanded. This program has the ability to generate every structure that could possibly be generated from the original drug structure, but incorporates template theory to selectively remove those structures which are unlikely to occur.