Efficient algorithm for quantitative assessment of similarities among atoms in molecules

Area of Science:

• Computational chemistry
• Quantum chemistry
• Molecular modeling

Background:

• Quantitative assessment of atomic similarity is crucial for understanding molecular properties.
• Accurate calculation of atomic similarity indices can be computationally intensive.
• Existing methods may not fully capture subtle variations in atomic environments.

Purpose of the Study:

• To develop a new, efficient algorithm for quantitatively assessing atomic similarity.
• To compute atomic similarity indices and their derivatives with respect to atomic orientation.
• To enhance the accuracy of similarity calculations without increasing computational cost.

Main Methods:

• Utilized recently developed analytical representations for atomic zero-flux surfaces.
• Developed an algorithm for efficient computation of atomic similarity indices.
• Calculated derivatives of similarity indices with respect to Euler angles for orientation analysis.

Main Results:

• The new algorithm efficiently computes atomic similarity indices and their derivatives.
• Enhanced accuracy in similarity index evaluation was achieved.
• The algorithm successfully discerned small changes in atomic similarity due to second-neighbor effects in carbonyl compounds.
• Demonstrated utility in detecting and quantifying steric interactions' effects on atomic shapes in acrolein.

Conclusions:

• The developed algorithm provides an accurate and efficient method for quantitative atomic similarity assessment.
• The approach enhances the understanding of electronic and steric effects on atomic properties within molecules.
• This method offers a valuable tool for detailed molecular analysis and computational chemistry research.