Multiblock modelling to assess the overall risk factors for a composite outcome.

Research in epidemiology may be concerned with assessing risk factors for complex health issues described by several variables. Moreover, epidemiological data are usually organized in several blocks of variables, consisting of a block of variables to be explained and a large number of explanatory variables organized in meaningful blocks. Usual statistical procedures such as generalized linear models do not allow the explanation of a multivariate outcome, such as a complex disease described by several variables, with a single model. Moreover, it is not easy to take account of the organization of explanatory variables into blocks. Here we propose an innovative method in the multiblock modelling framework, called multiblock redundancy analysis, which is designed to handle most specificities of complex epidemiological data. Overall indices and graphical displays associated with different interpretation levels are proposed. The interest and relevance of multiblock redundancy analysis is illustrated using a dataset pertaining to veterinary epidemiology.

Model cheese aroma perception is explained not only by in vivo aroma release but also by salivary composition and oral processing parameters.

The aim of the present paper was to determine, from four model cheeses differing in fat content and firmness and consumed by fourteen well characterised subjects, the respective impacts of in vivo aroma release, bolus rheology, chewing activity, mouth coating and salivary composition on dynamic aroma perception. The originality of the approach is that it considers all the parameters together and is able to evaluate their relative contribution using multi-block partial least square (MB-PLS) regression. The fruity aroma perception of the more hydrophilic compound (ethyl propanoate) was related to its dynamic release parameters before swallowing whereas the blue cheese aroma perception of the more hydrophobic compound (nonan-2-one) was related to its dynamic release parameters after swallowing and was highly impacted by mouth coating. Moreover the MB-PLS approach made it possible to evidence the combined effects of saliva composition and cross-modal interactions to understand why in some cases dynamic aroma perception could not be explained by dynamic in vivo aroma release data. Subjects with a low sodium content in saliva perceived fruity aroma which is not congruent with saltiness as less intense and salt- congruent (blue cheese) aroma as more intense, which was explained by their higher sensitivity to salt. Subjects with high lipolysis activity perceived fruity aroma which is not congruent to fat as less intense and fat-congruent (blue cheese) aroma as more intense, which should be explained by the link between lipolysis activity and fat sensitivity. These results could be considered for the reformulation of foods towards specific populations taking into account nutritional recommendations.