Fig. 3

a Predicted values for breakfast classification in ANOVA-Partial Least Squares (PLS) model. The data were ANOVA-decomposed into the factors Coffee/Tea, Gender, Individual and Breakfast type and PLS analysis was performed on the Breakfast type data after addition of the residual matrix. Model included 290 variables and 182 postprandial observations from 24 healthy volunteers. b Breakfast dependent cross-validated scores (cross-validated x scores (tcv)) in orthogonal projections to latent structures with discriminant analysis (OPLS-DA) model. Model included 290 variables and postprandial (3 h) serum samples (n = 182) from 24 healthy volunteers. c Predicted values in relation to response vector (Y) for volunteers in the in OPLS with effect projections (EP) model. The dotted line (Y = 1) indicates the response vector value for the model. The magnitude of the predicted effect for each volunteer is given by the height of the corresponding black bar. Deviations from the value 1 for a specific volunteer indicate a larger (> 1) or smaller (< 1) metabolic effect (difference between breakfast meals) in the model direction (metabolic profile) associated with the metabolism of foods included in the different breakfast meals. Model included 24 observations (equal to number of individuals) and 290 variables