Modeling gene regulation through end-to-end deep learning for expression prediction

During my PhD, I will explore the potential of sequence-based deep learning models to explore gene regulation. Current foundational deep learning models can predict chromatin accessibility and gene expression directly from the genome sequence for specific cell types. Using these models, I will perform in silico experiments to investigate the impact of enhancer and promoter sequence on accessibility and expression in different cell types. Furthermore, I aim to distill the knowledge of these large models into smaller, more focused models.