Precision fermentation of probiotic black rice: A digital twin framework integrating soft sensing and molecular co-pigmentation

Abstract

Black glutinous rice ( Oryza sativa L.) is a compelling functional beverage substrate due to its high cyanidin-3-glucoside (C3G) content. However, conventional fermentation faces an inherent trade-off: the extended incubation required for therapeutic probiotic titers (≥9.0 log CFU/mL) promotes acid-hydrolytic pigment degradation. This study establishes a precision fermentation framework using Lactiplantibacillus plantarum to resolve this viability–stability conflict. Response Surface Methodology (RSM) identified conditions (180 h, 20 °C) balancing cell viability (9.46 log CFU/mL) and anthocyanin retention (77.9%). For real-time monitoring, ethanol served as a robust soft sensor for viable cells ( r = 0.990). An Extended Kalman Filter (EKF) digital twin captured real-time dynamics, reducing the covariance trace by 83.1% upon measurement updates, yielding high accuracy (viability RMSE = 0.053 log CFU/mL). An inverse-RMSE-weighted ensemble forecaster predicted terminal plateaus 24 h ahead, which was augmented with transport-kinetic scaling factors for a hypothetical 1000 L industrial simulation. Bridging macro-metrics with molecular insights, computational docking provided a hypothesis-generating framework, suggesting enzymatically liberated ferulic acid theoretically stabilizes C3G via π-π stacking ( ΔG = −6.2 kcal/mol) to shield the C<inf>2</inf> position from hydration. Exhibiting high sequential reproducibility (CV ≤ 0.59%), this unified framework offers a scalable blueprint for precision food manufacturing.