AI-Model Integration for Predictive and Data-Driven Plant Systems Biology in Climate-Smart Agriculture

Abstract

The integration of artificial intelligence (AI) in plant system biology enables the understanding of plant organization, revealing complex relations at the biological level. Plant systems biology aims at resolving the complex networks that define growth, development, stress responses, and metabolism by reframing information of different levels of biological organization. With the increase in volume and complexity of these data, which includes genomics, transcriptomics, proteomics, metabolomics, and phenomics, conventional analytics is frequently finding itself unable to comprehend non-linear dynamics and high-dimensional data structures. The AI-based technologies, in particular, machine learning (ML) and deep learning (DL), provide a robust alternative. Supervised and unsupervised ML models are being utilized to unravel the gene regulatory networks, able to predict phenotype using genetic information, and identify biomarkers associated with the organism’s adaptation to the environment or disease resistance. Image-based phenotyping has made significant progress primarily through the application of DL models like convolutional neural networks (CNNs), and more recent ones, including recurrent neural networks (RNNs), are being adapted for sequence-based genomic analysis. AI plays an important role in the integration of multi-omics datasets like genomics, transcriptomics, proteomics, and metabolomics to develop comprehensive models for dissecting the complex genotype–phenotype relationships that govern plant adaptation. Furthermore, the application of generative AI, like large language models (LLMs), is being applied to biological datasets, enhancing the precision in predicting biological functions ranging from protein folding to gene regulation. The integration of AI with plant systems biology helps in the characterization of a diverse set of plant processes, thereby addressing key issues in crop improvement, environmental sustainability, and climate-smart agriculture. The timely progression of the computational tools is promising to make plant science more predictive and data-driven.