Biotech Today

Artificial intelligence (AI) is revolutionizing plant breeding by integrating machine learning (ML) and deep learning (DL) to enhance trait selection, genomic analysis, and crop improvement. AI-driven approaches enable high-throughput phenotyping, automated disease detection, and predictive breeding, improving efficiency and accuracy. Genomic selection (GS) and genome-wide association studies (GWAS) utilize AI to process high-dimensional genomic data, identifying SNP-trait associations and optimizing breeding programs. AI enhances image-based phenotyping through convolutional neural networks (CNNs) and computer vision for plant trait identification, stress analysis, and disease detection. Deep learning models, such as ResNet50, Inception V2, and EfficientNetV2-B4, have achieved over 90% accuracy in detecting crop diseases in bananas, maize, and wheat. AI-driven molecular breeding incorporates explainable AI (xAI) to improve GWAS model interpretability, addressing non-linear trait interactions and missing heritability. Integrating crop growth models (CGMs) with AI improves genotype-environment interaction predictions for traits like drought tolerance and yield. AI-based phenotyping platforms like CropQuant-Air use deep learning for wheat spike detection and yield classification, achieving over 97% accuracy. Automated machine learning (AutoML) tools, such as AutoKeras, enhance crop trait classification while reducing computational complexity. AI in genomic selection improves predictive accuracy by integrating molecular markers and environmental data, accelerating breeding cycles. AI-powered speed breeding and synthetic biology open new avenues for plant improvement, ensuring sustainable agriculture and food security.