Figure 6
Partial Least Squares-Multiblocs-Discriminant Analysis (PLS-MB-DA). PLS-MB-DA was used to determine what predictor variables, (i.e. biological determinants, stimulated saliva, resting saliva or microbiota = X blocks) were the most discriminant to characterize the subjects according to variables to be explained (BMI and/or orosensory sensitivity to lipids = Y blocks). (A) Comparison between NW (n = 21) and O (n = 17) subjects. (B) Comparison between lipid tasters (T, n = 26) and lipid non-tasters (NT, n = 12). (C) Comparison between obese tasters (OT, n = 9) and obese non-tasters (ONT, n = 9). Discriminant selection of variables was done using variable importance in the projection (VIP) with a threshold of 1. Mann & Whitney test was used to determine parameters that differed significantly between the three classes (NW vs O, T vs NT and OT vs ONT). Means ± SEM. §P < 0.1; *P < 0.05; **P < 0.01. Amyl, amylase; CA-VI, carbonic anhydrase-VI; Lipo, lipolysis; Lyso, lysozyme; LPS, lipoplolysaccharides; Prot, protein amount; TAC, total anti-oxidant capacity.
