Interest in class I MHC-mediated immunotherapy is growing rapidly. In order to fight a virus or cancer effectively, a successful immunotherapeutic must activate a large number of specific CD8+ T cells and also generate immunological memory. Attempts to generate immune responses towards tumor- or virus-derived peptides have frequently been frustrated by the nature of the peptide antigen itself. Either the peptide does not bind well to its cognate MHC, or the T cells directed towards it have been functionally inactivated in vivo. Altered-peptide ligands are an effective way to circumvent these problems. However, generating enhanced binding of altered peptides to class I MHC while still maintaining recognition of the wild-type peptide is not straightforward. Many groups design enhanced binding peptides by substituting the observed anchor residues with those that are most preferred by the class I MHC molecule. For many antigenic peptides, this approach does not work. Furthermore, if a higher affinity peptide is designed, the substitutions may result in reduced recognition by CD8+ T cells. Therefore, the design of an altered-peptide ligand requires careful testing of each candidate therapeutic in terms of affinity for class I MHC and immunological reactivity. Lastly, immunotherapy using class I MHC must also take into account the large genetic heterogeneity in the population. A therapeutic that is only effective for 5-10 percent of the population is not as attractive as one that works for over 90% of the population. The use of MHC supertypes (groups of class I MHC allotypes that share similar peptide-binding characteristics) shows great promise in overcoming this problem.