Machine Learning for Quantitative Prediction of Protein Adsorption on Well-Defined Polymer Brush Surfaces with Diverse Chemical Properties DOI

Shiwei Su,

Tsukuru Masuda, Madoka Takai

и другие.

Langmuir, Год журнала: 2025, Номер unknown

Опубликована: Март 12, 2025

Polymer informatics has attracted increasing attention because machine learning can establish quantitative structure-property relationships in polymer materials. Understanding and controlling protein adsorption on surfaces are crucial for various applications, such as immobilization supports, biosensors, antibiofouling surfaces. However, is a complex phenomenon that difficult to predict quantitatively owing the involvement of multiple factors. Therefore, this study aims model densely packed brushes with chemical structures, these well-suited analyzing correlations between polymer's structure amount during initial adsorption. Two proteins, bovine serum albumin (BSA) lysozyme, adopted target expectation differences their charge profiles will be reflected resulting model. The descriptors brush include grafted structures (thickness) properties, which described by contact angle ζ potential. This allows physicochemical knowledge incorporated into Random forest exhibits best performance all situations, accurately predicting amounts adsorbed BSA lysozyme. In addition, prediction potential also enables explainable based theoretical molecular descriptors, ensuring no characteristics overlooked. Moreover, used analyze contributions electrostatic hydrophobic interactions conclusion, developed surfaces, incorporating structure, angle, It provides predictions analyzes roles interactions, advancing design functional applications biosensors antifouling technologies.

Язык: Английский

Machine Learning for Quantitative Prediction of Protein Adsorption on Well-Defined Polymer Brush Surfaces with Diverse Chemical Properties DOI

Shiwei Su,

Tsukuru Masuda, Madoka Takai

и другие.

Langmuir, Год журнала: 2025, Номер unknown

Опубликована: Март 12, 2025

Polymer informatics has attracted increasing attention because machine learning can establish quantitative structure-property relationships in polymer materials. Understanding and controlling protein adsorption on surfaces are crucial for various applications, such as immobilization supports, biosensors, antibiofouling surfaces. However, is a complex phenomenon that difficult to predict quantitatively owing the involvement of multiple factors. Therefore, this study aims model densely packed brushes with chemical structures, these well-suited analyzing correlations between polymer's structure amount during initial adsorption. Two proteins, bovine serum albumin (BSA) lysozyme, adopted target expectation differences their charge profiles will be reflected resulting model. The descriptors brush include grafted structures (thickness) properties, which described by contact angle ζ potential. This allows physicochemical knowledge incorporated into Random forest exhibits best performance all situations, accurately predicting amounts adsorbed BSA lysozyme. In addition, prediction potential also enables explainable based theoretical molecular descriptors, ensuring no characteristics overlooked. Moreover, used analyze contributions electrostatic hydrophobic interactions conclusion, developed surfaces, incorporating structure, angle, It provides predictions analyzes roles interactions, advancing design functional applications biosensors antifouling technologies.

Язык: Английский

Процитировано

0