LAUSR

Abstract The present study sought to understand the correlation of the physicochemical properties of filtration membranes (FMs) with the global and the individual peptide migration during electrodialysis with filtration membranes (EDFM) using statistical approaches: redundancy (RDA) and multivariate regression analysis. Fourteen different FMs, characterized in terms of nine physicochemical properties, were tested during EDFM to simultaneously separate anionic and cationic peptides from a well-characterized complex whey protein hydrolysate. Based on RDA, four FM properties among nine: zeta potential, conductivity, roughness and percentage of macropores distribution, were significantly correlated with global peptide migration. Regarding individual peptide migration, in addition to aforementioned four FM properties, contact angle considerably affected the migration of at least one anionic and/or cationic peptide to respective recovery compartments. Furthermore, predictive multivariate regression models were also developed for migrated peptides based on the significant FM properties. A specific model estimates the migration behavior of a specific peptide, when FMs having a wide variety of physicochemical properties are used during EDFM. Furthermore, this study suggested that different interactions (electrostatic, size exclusion, hydrophilic/hydrophobic) occurring between FMs and peptides at their interface are triggered by physicochemical properties of both FMs and peptides, which significantly influenced global and individual peptide migration during EDFM.