Autoreactive T cells have been identified in most autoimmune diseases and recently even in healthy individuals. Similar, T cells that recognize either wild-type or tumorspecific tumor antigens have been increasingly reported to develop spontaneously in cancer patients. This insight has become possible mainly due to novel immunoassays which have revolutionized the discovery of rare antigen specific T cells. At present, the major dogma that explains this increasing number of reports of autoreactive T cells is that autoreactive T cells are counteracted by CD4+CD25+ regulatory T (Treg) cells in vivo, in particular in healthy individuals, whereas dysfunction in Tregs or Treg responsiveness may unmask the autoreactive T cell responses in patients with autoimmune diseases. However, studies that identify autoreactive T cells are usually performed by culturing T cells with antigen presenting cells loaded with E. coli produced recombinant protein or unmodified synthetic HLA binding peptides. Our concern is that this approach may ignore the presence of natural genetic variation and post-translational modifications such as e.g. the complex nature of N- and O-linked glycosylation of mammalian proteins. Thus, T cell antigen reactivities identified with unmodified antigens in vitro may in part represent in vitro T cell activation against neo-epitopes and not true in vivo autoreactivity as postulated. This methodological problem may have implications for the interpretation of the frequent reporting of autoreactive T cells in autoimmunity, T cell responses to wild-type tumor antigens in cancer patients and most important for the increasing reports on naïve T cells with specificity against self-antigens in healthy individuals. Here, we discuss and provide examples for the possibility that the experimental methodology applied to document T cell reactivity against unmodified protein or peptide may lead to overinterpretation of the reported frequencies of autoreactive CD4+ and CD8+ T cells.