Aromatase is a rate-limiting enzyme for estrogen biosynthesis that is overproduced in breast cancer tissue. To block the growth of breast tumors, aromatase inhibitors (AIs) are employed to bind and inhibit aromatase in order to lower the amount of estrogen produced in the body. Although a number of synthetic aromatase inhibitors have been released for clinical use in the treatment of hormone-receptor positive breast cancer, these inhibitors may lead to undesirable side effects (increased rash, diarrhea and vomiting; effects on the bone, brain and heart) and therefore the search for novel AIs continues. To meet this challenge, this research project seeks to employ various computational approaches (e.g. computational chemistry, molecular modeling, molecular docking and data mining) for the discovery and development of novel AIs spanning ligand-, structure- and systems-based approaches. During the course of the final year of this project, we have continued to further our understanding on the structure-activity relationship of compounds that can inhibit aromatase through various means of multivariate analysis including a method developed in-house called the Efficient Linear Method (ELM). Both ligand-based (e.g. quantitative structure-activity relationship) and structure-based (e.g. molecular docking) approaches were utilized to unravel the origin of aromatase and 17beta-hydroxysteroid dehydrogenase type 1 inhibitory activity on data sets ranging in size from small (less than 100 compounds) to large (almost 1,000 compounds) chemical libraries. Finally, we also utilized a chemogenomic approach known as proteochemometrics modeling to analyze the aromatase inhibitory activity data of a series of compounds against a series of aromatase variants in one unified predictive model, which is strike contrast to traditional QSAR modeling that considers only the ligands of interest and not the protein variants.