PROJECT SUMMARY Genome editing technologies have extraordinary potential as new genomic medicines that address underlying genetic causes of human disease; however, it remains challenging to predict their long-term safety, because we do not know the consequences of potential side effects of genome editing such as off-target mutations or immunogenicity. Our long-term goal is to understand and predict such unintended biological effects to advance the development of safe and effective therapies. T-cells are an ideal cellular model because: 1) they are highly relevant as the most widely used cells for development of therapeutic genome editing strategies (such as cell-based treatments for HIV and cancer) and 2) mature T-cells encode a diverse T-cell receptor repertoire that can be exploited as built-in cellular barcodes for quantifying clonal expansion or depletion in response to specific treatments. We, therefore, propose the following specific aims: 1) to predict which unintended editing sites have biological effects on human T-cells by integrating large-scale genome-wide activity and epigenomic profiles with state-of-the-art deep learning models and 2) to develop a human primary T-cell platform to detect functional effects of genome editing by measuring clonal representation, off-target mutation frequencies, immunogenicity, or gene expression. If successful, our experimental and predictive framework will profoundly increase confidence in the safety of the next generation of promising genome editing therapies.