Cancer relapse and therapy resistance in solid tumors are strongly shaped by immune reactions within the tumor microenvironment (TME). Our research focuses on how innate and adaptive immune programs determine tumor control, immune escape, and response to immunotherapy in ovarian cancer, pancreatic cancer, cholangiocarcinoma, glioblastoma, and supportive breast cancer models. A central theme is to understand how adaptive natural killer (aNK) cells, macrophages, B cells, stromal cells, and tumor cells interact in spatially organized tissue niches, and how these interactions can be exploited therapeutically. The project will use computational resources to analyze large multi-modal datasets generated from patient-derived tumor material, ascites, blood, tumoroids, and preclinical models. In high-grade serous ovarian cancer, we will define HLA-E-presented tumor peptide repertoires by integrating immunopeptidomics, transcript-guided proteogenomics, structural modeling, and functional NK-cell recall assays. We will identify tumor-derived HLA-E peptides that either suppress or reprogram aNK cells and use these data to build interpretable sequence–structure–function models for peptide-guided NK-cell immunotherapy. Previous work from the group has shown that adaptive NK cells can mediate tumor-specific immune memory and cytotoxicity in ovarian cancer, and that selected non-canonical HLA-E peptides can induce memory-like NK-cell responses. In parallel, we will analyze single-cell RNA-seq, spatial transcriptomics, multiplex imaging, and cytokine/proteomics data to define immune architectures in solid tumors. In pancreatic and biliary cancers, the project will focus on sex-dependent macrophage-centered immune states, systemic immune trajectories, and their association with immune exclusion, treatment response, and clinical outcome. The computational analyses will support the discovery of immune biomarkers, prioritization of therapeutic peptides and engineered aNK-cell strategies, integration of spatial and single-cell data, and development of predictive models for durable anti-tumor immunity. Overall, the project will generate mechanistic and translational knowledge to guide next-generation immunotherapies that improve treatment durability and reduce relapse in solid tumors.