SUMMARY Drug development of neurological disease and cancer relies heavily on our understanding of gene regulation. We know that thousands of regulatory factors in the DNA environment play a role in controlling gene expression. However, how these factors regulate gene expression across spatially and temporally continues to remain unclear. Testing all possible combinations, experimentally, will take an enormous amount of time and resources. Therefore, there is an urgent need to develop data-driven methods to extract relevant patterns from the experimental datasets. Existing deep learning-based studies can model the relationship between regulatory signals and gene expression. However, they fail to capture the underlying spatial and temporal structure of the data. Extracting this information is a challenging task that is critical to our understanding of how gene regulation drives cell development. Our research study aims to model the spatiotemporal gene regulation computationally. We will achieve this by integrating chromatin modification, three-dimensional interaction, regulatory signals, and gene expression data using deep learning methods. We plan to integrate 3D genome organization data with the existing regulatory element signals by using a graph-based neural network to predict gene expression. We will also implement neural network-based unsupervised methods that leverage the temporal structure in the existing gene expression and chromatin accessibility data to generate in silico gene expression measurements for new time-points. Integrating genome-wide spatial organization information and using temporal information will accurately highlight the factors related to gene regulation. By applying our methods to diseased and healthy cell lines, we will be able to identify the differentiating factors potentially involved in gene misregulation.