Docking Studies and High Throughput Virtual Screening for Breast Cancer Drug Designing Using a Distributed Environment

Abstract

Breast cancer is a cancer that forms in the tissues of the breast mainly in ducts and lobus, and this is considered as the second most common type of cancer. Estrogen Alpha and Estrogen Beta are the main protein receptors which increase the proliferation of Breast cancer cells by reacting with Estrogen hormone. Traditional drug designing process is very time consuming and very costly. Therefore, Docking and High Throughput Virtual Screening was used as the Computer Aided Drug Designing (CADD) method to reduce the overall time and the cost for drug designing process as much as possible. Estrogen Alpha was selected as the main cancer causing receptor and ZINC compound database with 13 million compounds and Sri Lankan indigenous plant data set with 200 chemical structures were used as the ligands to identify potential drugs for this Breast cancer dis- ease. Several time reduction and optimization processes were carried out to improve the e ciency of this CADD methodology. Hadoop Map-Reduce technology was used to run jobs in a parallel manner on a distributed environment. Compound data set was divided among machines by developing a mapper algorithm and Docking mechanism was integrated inside the reduce algorithm to run this Docking process on the cluster environment. Support Vector Machine active learning was used to identify the dockable and non-dockable compounds prior to the docking process. Finally we were able to achieve 87.25% of time reduction from overall drug designing process and we found 2 commercially available compounds and 6 indigenous plants which can be considered as the most potential drugs for this severe disease Breast cancer.