ABSTRACT Design and Delivery of Neoantigen-based Tumor Vaccines Neoantigens are a class of cancer antigen derived from mutations occurring within the genome of the tumor, and they present an attractive target for therapeutic tumor vaccines. Despite advancements in the development of neoantigen tumor vaccines, extensive challenges impede clinical translation. These challenges include how to optimally deliver neoantigenic peptides for vaccination and how to select for immunogenic epitopes from a set of computationally-predicted neoantigens. To address the challenges associated with neoantigen delivery (Aim 1), we will develop a poly(lactic-co- glycolic acid) (PLGA)-based neoantigen-delivering nanoparticle (ndNP) platform as a vehicle for vaccination in B16F10 melanoma and BBN963 basal bladder cancer models. We will study the interactions of these ndNPs with dual checkpoint-inhibitor/co-stimulator immunotherapy, measuring efficacy through tumor growth and survival studies. Immune monitoring through flow cytometry, RNA-sequencing, and IFN-γ ELISpot of tumor and draining lymph nodes will be used to study how therapeutic combinations differentially affect immune cell activation and phenotypic distributions. To address optimal selection of neoantigens for therapeutic vaccination (Aim 2), we will develop a predictive algorithm based on correlative analysis of intrinsic neoantigen features with their immunogenicity. Current neoantigen prediction methods have high false-positive rates, with no robust methods to predict for therapeutic efficacy of these predicted neoantigens. From the literature and preliminary studies, we believe five intrinsic neoantigen features correlate with immunogenicity: 1) predicted peptide/MHC binding affinity, 2) amino acid physicochemical characteristics, 3) mutational position, 4) gene expression, and 5) derivative gene function. Using B16F10 and BBN963 neoantigens that we previously predicted and validated with IFN-γ ELISpot, we will correlate these five intrinsic neoantigen features with ability to elicit IFN-γ production. Correlative analysis will include univariable and multivariable elastic net regression, and the predictive algorithm derived from these analyses will be validated using neoantigens predicted from the MC-38 colon adenocarcinoma tumor model. Pursuit of these aims could provide beneficial knowledge for the translation of neoantigen tumor vaccine strategies. The aims of this proposal will provide training in a unique skillset of immuno-oncology, nanotechnology, and computational biology. The training provided under this award will facilitate my development toward my career goal of leading an independently funded immuno-oncology laboratory focused on therapy development.