Conformational Pruning via the Permutation Invariant Root-Mean-Square Deviation of Atomic Positions

Area of Science:

• Computational Chemistry
• Structural Biology
• Cheminformatics

Background:

• The Cartesian root-mean-square deviation (RMSD) is crucial for comparing 3D molecular structures and conformations.
• Current RMSD methods face challenges with molecular symmetry and atom permutation, causing errors and inefficiency.
• Accurate structural comparison is vital for reducing computational costs in automated workflows.

Purpose of the Study:

• To define structural similarity for conformational ensembles.
• To develop an efficient algorithm for distinguishing molecular conformations.
• To address limitations of traditional RMSD in handling molecular symmetry and atom permutation.

Main Methods:

• Developed a permutation-invariant RMSD (iRMSD) approach.
• Incorporated a procedure for assigning canonical atom identities.
• Optimized atom-to-atom assignment for efficiency.

Main Results:

• The iRMSD method effectively handles symmetric molecules and multiple rotamers.
• Achieved significant reductions in computational complexity.
• Demonstrated suitability for large-scale conformational analysis.

Conclusions:

• The iRMSD approach provides accurate structural similarity assessment for molecular ensembles.
• Enables efficient pruning of duplicate conformations in computational chemistry.
• Facilitates cross-methodology ensemble comparison and automated property prediction.