Bitter-risk and masking
Targeting the compounds behind protein off-notes
High-protein beverages generate off-notes from their amino-acid content. A supervised classifier encodes each constituent as a molecular fingerprint and predicts a bitter-risk probability for every compound in the MoBai flavour universe. This points masking effort at the right targets; it does not predict consumer liking.
Bitter-risk screen
Each point is a compound positioned by molecular weight and predicted bitter probability across 79 compounds. Points above the threshold are flagged as bitter-risk. The branched-chain and aromatic amino acids dominate, consistent with the food-science literature on protein bitterness.
Five-mechanism masking strategy
Guided by the screen, five mechanisms are applied in the formulation. Together they target both the texture off-notes prominent in the English market and the aroma off-notes prominent in China, plus excessive sweetness.
Cyclic oligosaccharide encapsulation
Beta-cyclodextrin physically captures volatile off-note compounds within a molecular cavity, reducing the off-aroma at the receptors and directly addressing the off-smell complaint prominent in China.
Maillard reaction flavour masking
A controlled heat-induced reaction generates roasted and caramel notes that shift the flavour foreground away from protein off-notes.
Acidity and pH modulation
Mild acidification suppresses the bitter and astringent perceptions associated with protein hydrolysates.
Probiotic-driven fermentation
Fermentation with a documented probiotic strain contributes dairy-fresh notes that mask chalky and metallic perceptions, addressing the texture complaint prominent in the English market.
Tea-polyphenol matrix interaction
Polyphenols from the tea base bind protein molecules, reducing the free protein available to generate off-notes while contributing a controlled level of astringency.