Cloud-Enabled AI Solutions for Pathogen Pan-Genomics and Cultivar Selection in Precision Agriculture

Abstract

Machine Learning (ML) and Artificial Intelligence (AI) are revolutionizing pathogen pan-genomics and cultivar selection by leveraging cloud-based data solutions to enhance disease susceptibility prediction in crops. This study integrates Linear Regression, Lasso Regression, Decision Trees, Random Forest, Gradient Boosting, and XGBoost, optimizing their performance using Grid Search, Bayesian Optimization, and Genetic Algorithm (GA) in a cloud computing environment. Experimental results show that GA-optimized Lasso Regression achieved the lowest Mean Squared Error (MSE = 1.10) and the highest R² (-0.05), outperforming other models. Random Forest also demonstrated significant improvements (MSE reduced from 1.42 to 1.24), emphasizing the robustness of ensemble learning with evolutionary tuning. GA surpassed both Grid Search and Bayesian Optimization in efficiency and model generalization, showcasing its effectiveness in large-scale genomic data processing. This study highlights the potential of cloud-powered ML-driven genomic selection for disease resistance prediction, paving the way for optimized breeding strategies. Future research should explore deep learning, explainable AI (XAI), and real-time pathogen monitoring through cloud-based infrastructures to advance precision agriculture and sustainable crop management.