BACKGROUND This 4-center study is part of a project to validate a food allergy murine model for safety testing of hydrolyzed infant formulas. OBJECTIVE The aim of the current multi-center experiment was to evaluate the residual allergenicity of three partial hydrolyzed whey proteins (pWH) in a multiple-parameter cow's milk allergy murine model and to compare to the classically used guinea pig model. Previous work showed differences in the magnitude of the allergic response to whey between centers. To get a first insight in the effect of housing on the robustness of the mouse model, microbiota composition of non-sensitized mice was analyzed and compared between centers. METHODS Mice were sensitized intragastrically (i.g.) with whey, pWH or eWH using cholera toxin as an adjuvant. In mice, whey-IgE/IgG1, acute allergic symptoms were determined upon whey challenge. Guinea pigs were orally sensitized ad libitum via the drinking water (day 0-37) and challenged intravenously with whey on day 49. The microbial composition in fecal samples was determined in non-sensitized mice in all 4 research centers before and after conduct of the study. RESULTS Elevated levels of whey-IgG1 were detected in whey-sensitized mice in all centers. Except for pWH-A in center 4, we observed elevated levels of whey-IgE in whey-sensitized mice and mice sensitized with pWH-A, -B, -C. Center 2 was excluded from further analysis because of non-significant IgE levels in the positive control. In contrast to whey-mice, pWH-A treated mice showed no acute skin response, mMCP-1 release or change in body temperature upon whey challenge in all centers, which corresponds with the absence of anaphylactic shock symptoms in both the mouse and guinea pig model. pWH-B and -C induced anaphylactic shock symptoms in the guinea-pig and mice whereas results on the remaining allergic outcomes in mice were inconclusive. No differences in microbiota composition were measured in response to the challenge and Microbiota composition depended on the location of the centers. CONCLUSIONS Both animal models showed comparable results on the residual allergenicity of partial hydrolyzed whey proteins, but none of the centers was able to differentiate between the residual sensitizing capacities of the pWH-B and -C based on a single elicitation parameter in the murine model. Differences in microbiota composition might contribute to the robustness of the food allergy murine model. For a well-balanced prediction on the potential allergenicity of hydrolyzed infant formulas a multiple murine parameter model is suggested to decrease the risk of false positive or false negative results. A future challenge is to develop an overall scoring system for proper risk assessment, taking all parameters into account.