TOPSIS based multi criteria QSPR modeling of antibiotics using graph theoretic indices

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Cheminformatics

Background:

• Antibiotic drug discovery relies on understanding structure-property relationships.
• Topological descriptors offer a way to quantify molecular structure for predictive modeling.

Purpose of the Study:

• To investigate the structural features of antibiotic drugs using topological descriptors.
• To develop quantitative structure-property relationship (QSPR) models for predicting antibiotic properties.
• To rank antibiotics using computational descriptors and decision-making methods.

Main Methods:

• Utilized the M-polynomial framework to define topological descriptors.
• Calculated degree-based indices (e.g., Zagreb, Harmonic, Forgotten).
• Employed cubic and power regression models for QSPR, alongside TOPSIS and SAW for multi-criteria decision making.
• Applied entropy weighting for objective feature importance.

Main Results:

• Cubic regression models demonstrated higher prediction accuracy (larger determination coefficient, lower standard error) for QSPR.
• Integrated QSPR, graph theory, and decision analysis effectively ranked antibiotics.
• The entropy weighting scheme ensured objective feature importance.

Conclusions:

• The combined approach of QSPR modeling, graph theoretic indices, and entropy-based decision analysis is powerful for antibiotic screening.
• This methodology provides valuable insights for antibiotic drug discovery and optimization.
• Topological descriptors derived from the M-polynomial framework are effective molecular descriptors.