A framework for virtual screening

Recent advances in Bioinformatics and in Computer simulation and modelling have positively impacted the drug discovery process by turning viable the rational drug design (RDD). One of the major challenges in RDD is the understanding about protein-ligand interaction simulated at the atomic level by molecular docking algorithms. Virtual screening (VS) is defined as a computational approach applied to the analyses of large libraries of chemical structures in order to identify possible drug candidates to a target. The major challenge of VS based on molecular docking is the time required to run each experiment and the countless parameters and characteristics that should be defined by the researcher such as: the target(s) receptor, one or a set of ligands, the receptor binding site and so on. In order to perform more realistic docking simulations it is also necessary to account for the receptor and ligand flexibility. Therefore, this paper presents a framework for VS, where the user configure an experiment in a Web based platform informing the path of input and output files as well as the size, center and variation of the binding site(s). Then, the proposed framework generates a Python script that performs the VS experiment on the users personal computer. We expect that researchers from diverse backgrounds as Biology, Physics, Pharmacy, etc. can easily prepare VS experiments without the necessity of learning how to write scripts. To validate our proposed framework we performed five different case studies considering the AcrB protein as target receptor. All the case studies were easily realized using the proposed framework. The results show that the framework is effective to configure the VS experiments with different characteristics. Besides, the experiments can help on the search for new drug candidates for this important target.