A template-based automatic fragmentation algorithm for complex and large systems in the generalized energy-based...

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Molecular Modeling

Background:

• Energy-based fragmentation methods often require manual intervention, leading to inefficiencies and inconsistencies.
• Scaling quantum chemistry calculations to large and complex molecular systems remains a significant challenge.

Purpose of the Study:

• To develop a template-based automatic fragmentation algorithm to overcome limitations in existing methods.
• To extend the generalized energy-based fragmentation (GEBF) approach for efficient partitioning of diverse molecular systems.

Main Methods:

• Implemented a hierarchical SMILES-encoded GEBF template library for cyclic and acyclic functional groups.
• Utilized structure conversion, macrocycle detection, substructure matching, and small-fragment merging for partitioning.
• Controlled fragment sizes to balance accuracy and computational cost, with options for user-defined templates.

Main Results:

• Achieved quantum chemistry results comparable to conventional methods for various complex molecules (biomacromolecules, macrocycles, etc.).
• Reduced the largest subsystem basis size by over two-thirds, enabling practical computation of large systems (e.g., 1500-atom oligomers).
• Validated GEBF forces for accurate geometry optimizations, spectroscopic predictions, and reaction energy calculations.

Conclusions:

• The developed algorithm enables fully automated, scalable, and accurate quantum chemistry for large-scale applications.
• This advancement bridges the gap between theoretical chemistry and real-world applications by reducing computational cost.
• The method facilitates high-accuracy predictions for complex systems, including reaction mechanisms in enzymes.