This current project aims to develop innovative control strategies against powdery mildew, focusing on the interaction between melon (Cucumis melo) and Podosphaera xanthii, the causative agent. After understanding the biology and metabolic interactions of P. xanthii with its host, computer-aided molecular design, applying both supervised and unsupervised machine learning techniques are being used to discover new powdery mildew control agents, including natural compounds, peptides, and aptamers. These agents are being identified using advanced in silico methods to predict and design molecules with novel modes of action. The new potential agents will be tested experimentally, in order to confirm their in vitro and in vivo activity, aiming to develop new, effective agricultural treatments with significant potential for industrial application in crop protection. 

This research project focused on developing new phytoprotection tools targeting the cucurbit powdery mildew of Podosphaera xanthii as a model pathosystem, with a strong emphasis on machine learning and QSAR approaches. Initially, key proteins essential for P. xanthii development were identified through dsRNA-mediated gene silencing, revealing targets related to fundamental physiological processes. The project then leveraged QSAR based on molecular topology to identify acid phosphatase inhibitors, leading to the discovery of fungicidal compounds that inhibit phosphorus assimilation, validated through molecular docking and dynamics, with cantharidin and other new acid phosphatase inhibitors demonstrating strong efficacy in greenhouse trials.

This project focused on the development of chitin deacetylase inhibitors (CDAI) using QSAR and molecular docking approaches for controlling powdery mildew, specifically targeting Podosphaera xanthii in cucurbits. Through advanced chemoinformatics techniques, a collection of CDAIs was designed and subjected to three rounds of virtual screening. The top candidates demonstrated significant fungicidal activity in initial assays on zucchini cotyledon discs against P. xanthii. Further testing in melon plants showed a 90% reduction in disease, with similar efficacy against Botrytis cinerea in tomato fruits. The CDAIs' mechanism of action was confirmed via molecular docking and enzymatic inhibition assays, supported by experiments on melon seedlings with silenced CmCERK1, where fungal growth was restored, indicating successful chitin signaling disruption. Though greenhouse trials with these inhibitors were limited, they provide a promising foundation for future fungicide development.

Given the lack of current effective treatments against Xylella fastidiosa, the search for substances capable of eliminating or neutralizing this bacterium is a top priority. To address this challenge, computational methods based on QSAR and molecular topology were applied for the design of an effective in silico strategy to design new Xylella fastidiosa control tools.