Predicting intensity interactions in complex odor mixtures

Natural odors are complex mixtures with typically 10-30 components drawn possibly from over 10,000 odorants. In mixtures, odor molecules compete, overshadow, suppress, inhibit and synergize with each other. For an accurate prediction of the intensity of odor mixture, the type and prevalence of mixture interactions need to be known. To identify mixture interactions, human observers rated the intensity of individual odors and their mixtures. Twenty-four individual odorants (dose-response at 7 concentrations) and subsequent 2, 3, 5, and 10 component mixtures were evaluated. Biophysical models were developed to predict human observer response to odor mixtures using the response to individual odorants. Widely used odor mixture models such as linear model, additive model, and vector model significantly over-predicted mixture response. Biophysically plausible models that account for molecular interactions at the receptor levels such as competitive binding of molecules to receptors or occupancy of receptors by the strongest component provided a closer match to experiments.