In the advent of a rapidly changing climate, there is a need for rapid selection and editing of plants to develop climate resilience. To aid in these efforts, interpretable and accurate mechanistic models of gene regulation will need to be developed in a multitude of species. This dissertation presents a comprehensive exploration of computational techniques for predicting gene expression in maize, focusing on regulatory network-based machine learning and deep learning approaches. These models have significant implications for gene editing in crop plants, providing information on the regulatory mechanisms that govern gene expression. The first section investigates the integration of regulatory network information into machine learning models to predict gene expression. Through extensive experimentation, various network-based approaches are evaluated, with the aim of improving prediction accuracy by capturing the complex shared regulatory mechanisms that underlie gene expression in maize. The next study focuses on the development of PLExBench, a benchmark suite specifically tailored for predicting gene expression in plants. PLExBench encompasses multiple tasks in model plants of Arabidopsis thaliana and Zea mays, serving as a standardized platform to evaluate the performance of gene expression prediction methods. By rigorously evaluating state-of-the-art prediction algorithms using PLExBench, this work introduces an effective strategy to assess the strengths and limitations of existing methods in accurately predicting gene expression in plants. Finally, this dissertation explores the application of deep learning techniques to uncover cis-regulatory contributions to tissue-specific gene expression in maize. Using perturbation-driven experimental data, deep learning models are used to decipher the intricate regulatory mechanisms underlying tissue-specific gene expression patterns. Through systematic analysis and validation, the study elucidates the capabilities of deep learning approaches in capturing cis-regulatory elements and predicting tissue-specific gene expression profiles in maize. By combining insights from regulatory network-based machine learning, benchmarking efforts with PLExBench, and perturbation-driven assessment of deep learning, this dissertation contributes to a deeper understanding of the use of deep learning to find regulatory components of gene expression. The findings provide valuable resources and methodologies to improve gene expression prediction models in agricultural contexts and advance our understanding of plant biology, thus facilitating targeted gene editing efforts to enhance crop traits and agricultural productivity.