Our long-term goals are to uncover general principles of high-order interactions in the human dark genome, and to elucidate mechanisms by which genetic variants impact human physiology and health. We use a combination of interdisciplinary approaches, including multiplexed CRISPR screening, deep learning modeling, and statistical methods to uncover interactive logics within the non-coding regulatory networks and reveal the DNA sequence codes in governing these genomic systems. We are interested in use these rules and codes to probe the causal consequences of genetic mutations on complex disease and develop novel AI-driven targeted therapy. We aim to lead our research towards clinical applications for early diagnoses and targeted therapy in precision cancer medicine.