Algorithm for Efficient Superposition and Clustering of Molecular Assemblies Using the Branch-and-Bound Method

Area of Science:

• Computational chemistry
• Structural bioinformatics
• Cheminformatics

Background:

• Root-mean-square deviation (RMSD) is crucial for comparing 3D chemical structures, impacting predictions of reactivity, properties, and bioactivity.
• Determining atom mapping and spatial superposition for RMSD calculation is computationally challenging, especially for large systems.

Purpose of the Study:

• To introduce mobbRMSD, a novel algorithm for exact RMSD calculation that addresses limitations of existing methods.
• To enable efficient structural similarity analysis for complex chemical systems and molecular dynamics trajectories.

Main Methods:

• Formulation in molecular-oriented coordinates and application of the branch-and-bound method for exact RMSD solutions.
• Incorporation of chemical knowledge (atom types, bonding, chirality) for handling diverse chemical systems.
• Development of a mobbRMSD-based structural clustering method for molecular dynamics data.

Main Results:

• mobbRMSD extends the system size limit for exact RMSD solutions by nearly double compared to previous methods.
• Successfully analyzed structural similarity in large molecular micelles, a task difficult for prior algorithms.
• Achieved asymptotically average polynomial time complexity for structural clustering of molecular dynamics trajectories.

Conclusions:

• mobbRMSD provides an efficient and exact solution for RMSD calculation, advancing structural similarity analysis in computational chemistry.
• The algorithm enhances the study of molecular liquids, solute solvation, self-assembly, and molecular dynamics simulations.
• mobbRMSD offers a significant improvement for analyzing large molecular systems and trajectory data.