AI-Assisted Genetic Clustering Assessment of Date Palm (Phoenix dactylifera L.) Cultivars Through Mitochondrial DNA RAPD Markers

Abstract

Understanding the genetic diversity of date palms is essential for improving and conservation of cultivars. In this study, we combined RAPD-PCR techniques with AI-powered data analysis to explore the genetic diversity of 15 date palm cultivars involving dry, semi-dry, and soft kinds. Using mtDNA as the template, ten random primers were employed to amplify DNA fragments, and AI algorithms assisted in band detection, binary matrix generation, and similarity index calculations. The integration of AI enabled accurate clustering (using Jaccard and Dice coefficients), identification of highly informative primers (notably OPA9), and pattern recognition in PCA space. Hierarchical clustering and PCA were performed for genetic relationships assesment. This human–machine collaboration revealed hidden genetic relationships and confirmed divergence between cultivars like Barhee and Ghazal. The study highlights how AI tools can elevate classical genetic fingerprinting, providing deeper insights for breeding, conservation, and cultivar management. The results revealed high polymorphism with notable discriminatory power by primers OPA9, OPA11, and OPA12. Hierarchical clustering grouped cultivars into three main genetic clusters, independent of fruit type, while PCA further clarified inter-cultivar relationships. The identification of genetically distinct cultivars like Samany and Ghazal provides valuable targets for genetic preservation, while close pairs like Hayany and Zaghloul suggest possible clonal or local lineage relationships.