Predicting Adhesive Free Energies of Polymer--Surface Interactions with Machine Learning

Abstract

Polymer-surface interactions are crucial to many biological processes and industrial applications. Here we propose a machine-learning method to connect a model polymer's sequence with its adhesion to decorated surfaces. We simulate the adhesive free energies of 20, 000 unique coarse-grained 1D sequential polymers interacting with functionalized surfaces and build support vector regression (SVR) models that demonstrate inexpensive and reliable prediction of the adhesive free energy as a function of the sequence. Our work highlights the promising integration of coarse-grained simulation with data-driven machine learning methods for the design of new functional polymers and represents an important step toward linking polymer compositions with polymer-surface interactions.

Predicting aggregate morphology of sequence-defined macromolecules with recurrent neural networks

“…67,68 Previous applications of ML to materials design have explored features based on combinations of thermodynamic, chemical, and topological information to manually create engineered features, 69 representing macromolecules as chemistry-informed graph based features, 70 converting monomeric sequences to image-based features, 71 or simple one-hot encoded features. 72 In this work, we apply supervised ML to predict the aggregation behavior of a model copolymer. Motivated by the variety of featurization techniques described in the literature, we consider three different encoding schemes to encode the monomer sequences that require information about only the monomer arrangement.…”

Machine learning strategies for the structure-property relationship of copolymers

“…Based on random copolymers and homopolymers, Hanaoka (2020) , Leibfarth et al., ( Reis et al., 2021 ), Kosuri et al. (2022) , Shi et al. (2021) , Tamasi et al.…”