Ali Bashashati

This study uses GANs to synthesize high-resolution cancer pathology images, showing they are visually indistinguishable from real images and perform as well as real images in diagnostic tasks.

This study introduces a machine learning method to infer mutation signatures from both single nucleotide and structural variations in DNA repair deficient cancers, identifying new age-related and HRD-associated signatures in breast and ovarian cancers.

The paper analyzes the interaction between tumor and immune clones in ovarian cancer, finding that immune contexture patterns mirror malignant spread and that foldback inversions correlate with poor outcomes.