AI-Integrated Omics Analysis Reveals Cultivar-Specific Resistance Mechanisms to Powdery Mildew in Cucurbita pepo

Powdery mildew represents one of the most significant challenges for cucurbit crops. In recent decades, progress has been made in identifying resistance sources that improve yield and quality while reducing fungicide use. This study explored the molecular mechanisms underlying cucurbit responses to powdery mildew through comparative RNA-seq of two contrasting <i>Cucurbita pepo</i> cultivars: the partially resistant 968Rb and the susceptible True French. Differential expression analysis between inoculated and non-inoculated conditions identified 398 DEGs in 968Rb and 1129 in True French. In 968Rb, a stronger defense response emerged with cell wall reinforcement and upregulation of <i>fructose-1,6-biphosphate aldolase</i> genes, while True French showed activation of chitinase genes. Machine learning models, including Random Forest and K-means, identified expression features and gene modules linked to resistance. By combining conventional and Artificial Intelligence-based analyses, we identified a putative adaptive genetic variation, supported by a higher single nucleotide polymorphism density within expression clusters enriched for upregulated genes in the partial resistant cultivar 968Rb. The integration of Artificial Intelligence tools in our pipeline facilitated the understanding of the genetic basis of <i>Cucurbita pepo</i> resistance to <i>Podosphaera xanthii</i>, highlighting the transcriptional modules and variant patterns associated with resistance traits, and providing a scalable framework for future applications in crop improvement.