Evaluation of antiarrhythmia drug through QSPR modeling and multi criteria decision analysis

Abstract

This study explores how topological indices (TIs), which are mathematical descriptors of a drug's molecular structure, can support to predict vital properties and biological activities. This understanding is a key for more effective drug design. We focused on drugs used to treat several arrhythmia conditions, including tachycardias, bradycardias, and premature beats. Our approach combines molecular modeling with decision-making techniques to offer a cost-effective way to understand how these drug molecules behave. Our procedure started with calculating topological indices for the chemical structures of these medications to extract information about their features. We then established quantitative structure-property relationship (QSPR) models using quadratic regression, training and validating them. We concentrated on TIs that showed a strong correlation[Formula: see text] with physicochemical properties. Each property was also weighted, based on its correlation with the topological indices. As a final point, to aid in informed decision-making, we employed multiple-criteria decision-making approaches Technique for Order Preference by Similarity to Ideal Solution TOPSIS and Simple Additive Weighting SAW to rank the anti- arrhythmia medications. Drug Amiodarone ranked highest due to strong correlation with boiling point and polarizability. The study also highlights the potential of machine learning to analyze large datasets, allowing for accurate predictions of chemical behavior. This comprehensive method can facilitate the detection of new drugs with valuable qualities and improve our understanding of how chemical structures affect drug effectiveness.

Keywords: Correlation coefficient; Heatmap; MCDM; QSPR analysis; SAW; TOPSIS; Topological indices.

Fig. 1

Structures of anti-arrhythmia drugs.

Fig. 2

Methodology.

Fig. 3

Heatmap of topological indices vs physicochemical properties.

Fig. 4

Plots of quadratic regression models that predict optimal physicochemical properties using degree-based topological indices.

Fig. 5

Weights allocated to topological indices.