On Relationships of Eigenvalue-Based Topological Molecular Descriptors.

Area of Science:

• Cheminformatics
• Quantitative Structure-Property Relationships (QSPR)
• Quantitative Structure-Activity Relationships (QSAR)

Background:

• Topological molecular descriptors are crucial for QSPR/QSAR studies.
• Eigenvalue-based descriptors offer unique insights into molecular structure.
• The predictive power of novel descriptors requires rigorous validation.

Purpose of the Study:

• To compare three eigenvalue-based topological molecular descriptors.
• To investigate the relationships between these descriptors.
• To identify structural parameters influencing these relationships.

Main Methods:

• Utilized multiple datasets of alkanes.
• Applied multiple linear regression (MLR) analysis.
• Compared two established descriptors with a newly derived one.

Main Results:

• Established relationships among the three investigated molecular descriptors.
• Identified key structural parameters governing these relationships.
• Developed MLR formulas to describe descriptor interrelations.
• Demonstrated that all three descriptors capture largely similar molecular information.

Conclusions:

• The three eigenvalue-based topological descriptors are highly correlated.
• Differences lie in their sensitivity to molecular branching and non-bonding orbitals.
• The newly derived descriptor shows comparable information content to established ones.