The current methodology to identify allergenic food proteins is effective in identifying those that are likely to cross-react with known allergens. However, most assays show false positive results for low/non-allergens. Therefore, an ex vivo/in vitro DC-T cell assay and an in vivo mouse model were used to distinguish known allergenic food proteins (Ara h 1, -Lactoglobulin, Pan b 1, bovine serum albumin, whey protein isolate) from low/non allergenic food proteins (soy lipoxygenase, gelatin, beef tropomyosin, rubisco, Sola t 1). CD4+ T cells from protein/alum-immunized mice were incubated with corresponding protein-pulsed bone marrow-derived DC and analyzed for cytokine release. All known allergens induced Th2 responses in vitro, whereas soy lipoxygenase, gelatin or beef tropomyosin did not. Sola t 1 and rubisco induced a more generalized T cell response due to endotoxin contamination, indicating the endotoxin-sensitivity of the DC-T assay. To analyze responses in vivo, mice were orally sensitized on days 0 and 7. Known allergens induced IgE and mMCP-1 release upon oral challenge at day 16, whereas the low/non-allergens did not. Both the DC-T cell assay and the mouse model were able to distinguish 5 known allergens from 5 low/non-allergens and may be useful to identify novel allergenic food proteins.