ARTIFICIAL INTELLIGENCE APPLIED TO VIRTUAL SCREENING FOR CONSENSUS SCORING

The occurrence of pandemics and outbreaks is a recurring phenomenon throughout human history, exacerbated by the increasing ease of global mobility, which accelerates the spread of microorganisms. To mitigate these effects, sanitary measures are adopted, along with the pursuit of effective treatments. However, drug development is an expensive and time-consuming process, making it essential to apply techniques that can accelerate and reduce its costs. A promising approach in this context is the use of artificial intelligence (AI) in rational drug design. This study proposes the development, implementation, and validation of a molecular docking system based on consensus scoring, using AI to assist in the virtual screening of compounds with therapeutic potential. The methodology involves training a predictive model based on molecular descriptors of ligands and receptors, extracted from specialized databases, combined with docking scores obtained from different software programs. The central hypothesis is that the consensus-based model will have a greater ability to distinguish biologically active compounds from inactive ones, considering not only the individual results of the docking programs but also the factors that influence their variation. The expected outcomes include improved efficiency in selecting promising compounds for drug development, contributing to the optimization of the drug discovery process and accelerating the response to outbreaks and pandemics.