Novel framework for predicting multitarget antifungal drugs against pathogenic fungi

Abstract

Pathogenic fungal diseases have become a global threat to human health in immunocompromised individuals. With increasing mortality rates due to systemic infections and limited availability of antifungal classes, it is crucial to accelerate antifungal drug discovery by focusing on novel targets. In our research, we aim to target Rho proteins and develop a deep-learning framework for antifungal drug prediction. Stable structures of plasma membrane conserved Rho proteins of different pathogenic species were generated using ESMFold (Evolutionary scale modeling). Our investigation focused on establishing cavity-specific in silico ligand synthesis with a further emphasis on exploring orthologous cavity sites across pathogenic species. To assess the compounds generated via LigBuilder we applied rigorous statistical methods to segregate them into high and low-affinity binders. Subsequently, we employed graph neural networks for the evaluation. In addition to targeting novel pathway, we explored drug repurposing as an avenue for therapeutic alternatives against infections. Leveraging the Drug Repurposing Hub for prediction, we incorporated repurposed drugs into our research and conducted docking studies to validate their potential for experimental studies. This approach adds a valuable dimension to our investigation, aligning with the growing interest in repurposing drugs for infections.