Artificial Intelligence Models Bridging Genotype, Environment and Phenotype in Agricultural Genomics

Abstract

Genotype–environment–phenotype (GEP) interactions are central to agricultural productivity, resilience, and sustainability. Genotype determines how organisms respond to environmental variability, shaping phenotypic outcomes such as yield, quality, and stress tolerance. However, the complex, nonlinear, and stochastic nature of genotype–environment (GE) interactions poses major challenges for prediction and decision-making in breeding and agronomic management. Recent advances in artificial intelligence (AI) offer powerful tools to bridge genotype, environment, and phenotype by integrating heterogeneous data sources, modeling high-dimensional interactions, and enabling predictive and inferential analyses across biological scales. This work presents a conceptual and methodological framework for AI-driven GEP modeling in agricultural genomics, encompassing data acquisition, preprocessing, and integration of multi-omics, environmental sensing, phenotyping platforms, and management metadata. We highlight the role of supervised learning and related AI approaches in learning GE relationships, predicting phenotypic outcomes under specific environmental conditions, and supporting genotype selection and deployment strategies. By moving beyond descriptive associations toward causal understanding, AI-enabled GEP frameworks can enhance breeding efficiency, optimize resource use, and improve the adaptability of crops and livestock to diverse and changing environments