Predicting tuberculosis drug properties using extended energy based topological indices via a python driven QSPR...

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Tuberculosis (TB) remains a significant global health challenge, necessitating continuous development of effective treatments.
• Understanding the physicochemical properties of anti-TB drugs is crucial for optimizing their efficacy and safety.

Purpose of the Study:

• To explore physicochemical characteristics of key anti-TB drugs using extended energy-based topological indexes.
• To develop predictive models for essential physicochemical parameters of anti-TB medications.
• To assess the utility of a Python-based Quantitative Structure-Property Relationship (QSPR) methodology in drug design.

Main Methods:

• Calculation of extended energies using topological indexes (Zagreb Second, Harmonic, Randic, Sombor, Reduced Sombor, Average Sombor) for anti-TB drugs.
• Application of linear, quadratic, and logarithmic regression models to predict nine physicochemical parameters.
• Utilizing Python for algorithmic implementation, matrix formulation, and eigenvalue computation for reproducibility.

Main Results:

• Quadratic regression models consistently provided the best predictability for physicochemical properties.
• High correlations were observed between topological descriptors and drug properties.
• Visual analyses (heatmaps, scatter plots, bar charts) supported the numerical findings.

Conclusions:

• The Python-based QSPR methodology effectively correlates topological descriptors with drug properties.
• This approach offers an efficient pathway for drug design and optimization in pharmaceutical research.
• The study emphasizes transparency and reproducibility through publicly available code and data.