T cell receptor (TCR)-based therapeutic cells and agents have emerged as a new class of effective cancer therapeutics against intracellular cancer-associated proteins. These agents rely on presentation of short peptides derived from cellular, viral or phagocytosed proteins on major histocompatibility complex (MHC). However, cross-reactivities of these agents to off-target cells and tissues are poorly understood, difficult to predict, and have resulted in serious, sometimes fatal, adverse events. We have developed a mammalian, minigene-based method (termed “PresentER”) that encodes MHC-I peptide ligands for functional immunological assays as well as for determining the reactivities and potential cross-reactivities of TCR-like therapeutic agents against libraries of MHC-I ligands. This system is highly specific to, and entirely dependent on, the genetically encoded MHC peptide sequence, because it does not require proteasome cleavage, transporter associated with antigen processing (TAP) or processing, for immune presentation. Cells expressing PresentER antigens can be bound by TCR and TCR mimic (TCRm) antibodies, activate antigen-specific T cells, lead to antigen-specific cell death in vitro and tumor rejection in vivo. Using PresentER in a pooled library screen, we find dozens of MHC-I ligands encoded in the human proteome that are cross-reactive with two TCR mimic antibodies and are not predictable by other methods. We extend the use of this method to find the targets of an engineered TCR. Finally, we leverage the ability to generate tumors comprised of libraries of MHC-I ligands to study the determinants of MHC-I peptide immunogenicity in vivo. Surprisingly, we show that highly immunogenic tumor antigens (encoded by PresentER) do not lead to early immune-mediated tumor rejection in vivo when the fraction of cells bearing each antigen (“clonal fraction”) is low. Moreover, the clonal fraction that must bear an antigen in order to lead to rejection of immunogenic tumor subclones is dependent on the individual antigen itself. These data indicate that tumor neoantigen heterogeneity has an underappreciated impact on the ability of the immune system to detect and eliminate cancer cells and has implications for the prioritization of antigens in the design of novel antigen-specific immunotherapeutics such as cancer vaccines.