Quantitative structure-property relationships generated with optimizable even/odd Wiener polynomial descriptors

Area of Science:

• * Cheminformatics
• * Computational Chemistry
• * Physical Chemistry

Background:

• * Molecular structure is numerically characterized by structural descriptors, such as the Wiener index (W).
• * Distance-based topological indices are crucial in Quantitative Structure-Property Relationship (QSPR) and Quantitative Structure-Activity Relationship (QSAR) models for drug design, toxicology, and virtual screening.
• * The Wiener index (W) is a foundational descriptor derived from interatomic distances in molecular graphs.

Purpose of the Study:

• * To introduce novel topological indices derived from partitioning the Wiener polynomial.
• * To explore the utility of these novel indices in enhancing QSPR models.
• * To investigate the predictive power of even and odd molecular graph distance counts.

Main Methods:

• * Development of novel topological indices by partitioning the Wiener polynomial based on even and odd molecular graph distances.
• * Optimization of even and odd power function variables within QSPR modeling.
• * Application and testing of these indices in QSPR models for various physical properties of alkanes.

Main Results:

• * The proposed even/odd Wiener polynomial indices demonstrate improved correlations for predicting physical properties of alkanes.
• * These novel indices can lead to simpler and more effective QSPR models.
• * Notable improvements were observed in correlations for boiling temperature, molar heat capacity, standard Gibbs energy of formation, vaporization enthalpy, refractive index, and density.

Conclusions:

• * Novel topological indices based on even and odd Wiener polynomial partitioning offer enhanced predictive capabilities for QSPR.
• * These indices provide a more refined approach to molecular descriptor development.
• * The findings suggest a promising avenue for improving the accuracy and simplicity of QSPR models in cheminformatics.