Abstract
This webinar is based on Albert and colleagues recent study exploring the use of AlphaFold-predicted protein structures for drug discovery against Chagas disease, a neglected tropical disease caused by the parasite Trypanosoma cruzi. The study used AlphaFold-predicted models of T. cruzi proteins to carry out a structure-based virtual screening of approximately 30,000 compounds. From this screening, top-ranked candidates were selected for experimental testing, and two already approved drugs showed significant specific antiparasitic activity. These results illustrated how AlphaFold-based virtual screening can support drug discovery efforts and accelerate the identification of new compounds against neglected tropical diseases, where the absence of experimentally determined structures is problematic. Building on this experience, the webinar will walk through how AlphaFold can be applied to structure-based drug discovery more broadly, using Chagas disease study as an example. It will briefly introduce AlphaFold, covering its confidence metrics and where to download or generate predicted models. From there, the webinar will show the steps needed for running a virtual screening campaign: predicting and assessing binding pockets, downloading chemical libraries, tools for molecular docking, and processing and interpreting the resulting data. Throughout, a particular attention will be given to the caveats and limitations of using predicted rather than experimentally determined structures for docking, drawing on the challenges encountered in their study. The webinar will close by discussing what to keep in mind when adapting this workflow to other cases.