Breeding vegetables for whitefly resistance: past, present, and future in the AI era

Whiteflies, particularly <i>Bemisia tabaci</i>-a rapidly evolving cryptic species complex comprising more than 40 biotypes including the invasive MEAM1 and MED-and <i>Trialeurodes vaporariorum</i>, remain among the most destructive pests of global vegetable production. Their adaptability, wide host range, and efficient virus transmission drive recurrent epidemics in crops such as tomato, pepper, eggplant, cucurbits, and snapbean. Over six decades, breeding for whitefly resistance has progressed from phenotypic selection to the identification of resistance mechanisms such as antibiosis, antixenosis, and tolerance, and to the exploitation of diverse sources from wild relatives and landraces. Recent advances in QTL mapping, pangenomics, multi-omics integration, genomic selection, and CRISPR-based modification of metabolic and structural defense traits have transformed the landscape of resistance breeding. Emerging AI-enabled tools-including machine-learning models for automated whitefly phenotype detection, hyperspectral stress diagnostics, and predictive modelling of resistance loci-are accelerating the dissection and deployment of complex traits. Importantly, durable whitefly resistance enhances climate resilience by reducing dependence on insecticides, stabilizing yields under abiotic-biotic stress combinations, and mitigating climate-driven surges in whitefly populations and virus epidemics. By integrating classical genetics, modern biotechnology, multi-omics, and AI-driven decision frameworks, breeding programs can more rapidly develop robust, climate-resilient vegetable cultivars capable of withstanding evolving whitefly threats.