Developing stacking models for interacting biopolymers using quantum mechanical calculations and machine learning

Abstract

Biopolymers are long-chain molecules with significant potential for developing sustainable, renewable, and environmentally friendly materials. Agarose, a biopolymer derived from red algae, is widely used in bioengineering and biomedical applications because of its biocompatibility and biodegradability. Understanding how agarose chains interact and assemble is essential for explaining its gelation behavior and material properties. For decades, scientists have debated whether agarose chains primarily form double-helical structures or other arrangements. In this study, researchers used high-throughput quantum mechanical calculations and machine-learning models to investigate the interactions between agarose chains. Their results show that stacked configurations are energetically and statistically more favorable than the traditionally proposed double-helix structure. The study also identifies simple interatomic interactions and distances between specific atomic species as effective descriptors for predicting the assembled structures. These findings resolve a long-standing controversy and provide valuable insight into agarose structure, gelation, and the design of advanced biopolymeric materials.