Systematic modification of functionality in disordered elastic networks through free energy surface tailoring

Area of Science:

• Computational chemistry
• Materials science
• Machine learning

Background:

• Molecular and materials engineering requires understanding complex system dynamics.
• Traditional methods often struggle with extensive connectivity and intricate interactions.
• Predictive modeling is crucial for designing materials with specific functionalities.

Purpose of the Study:

• To develop a hybrid machine learning-physics approach for molecular and materials engineering.
• To identify and modulate critical molecular interactions for system tailoring.
• To demonstrate the approach's efficacy in engineering allosteric regulation and strain fluctuations.

Main Methods:

• Constructing collective variables using machine learning models trained on system-specific data.
• Applying these collective variables to analyze and modify free energy landscapes.
• Utilizing enhanced sampling simulation concepts.

Main Results:

• Successfully identified critical molecular interactions within a complex disordered elastic network.
• Demonstrated the ability to systematically tailor the system's free energy landscape.
• Engineered allosteric regulation and uniaxial strain fluctuations effectively.

Conclusions:

• The combined approach offers a powerful tool for molecular and materials design.
• Provides insights into functionality governed by extensive connectivity.
• Highlights potential for designing complex molecular systems with tailored properties.