Machine-Learning-Aided Understanding of Protein Adsorption on Zwitterionic Polymer Brushes

Abstract

Constructing antifouling surfaces is a crucial technique for optimizing the performance of devices such as water treatment membranes and medical devices in practical environments. These surfaces are achieved by modification with hydrophilic polymers. Notably, zwitterionic (ZI) polymers have attracted considerable interest because of their ability to form a robust hydration layer and inhibit the adsorption of foulants. However, the importance of the molecular weight and density of the ZI polymer on the antifouling property is partially understood, and the surface design still retains an empirical flavor. Herein, we individually assessed the influence of the molecular weight and density of the ZI polymer on protein adsorption through machine learning. The results corroborated that protein adsorption is more strongly influenced by density than by molecular weight. Furthermore, the distribution of predicted protein adsorption against molecular weight and polymer density enabled us to determine conditions that enhanced (or weaken) antifouling. The relevance of this prediction method was also demonstrated by estimating the protein adsorption over a wide range of ionic strengths. Overall, this machine-learning-based approach is expected to contribute as a tool for the optimized functionalization of materials, extending beyond the applications of ZI polymer brushes.