Structural analysis of anti-cancer drug compounds using distance-based molecular descriptors and regression models

Abstract

Molecular descriptors encapsulate the key structural information of molecules, which is crucial for elucidating molecular behaviors. They have proven invaluable in quantitative structure–property relationship (QSPR) analysis. Such studies involve rigorous scientific investigations into the relationship between molecular structure and diverse physicochemical properties, revealing the underlying principles governing structure–property correlations. This facilitates predictive modeling and rational design across a wide range of scientific disciplines. Cancer is a lethal disease characterized by the uncontrolled growth and spread of abnormal cells. This study aims to develop regression models for predicting physicochemical properties of novel anti-cancer drugs targeting blood and skin cancers. Utilizing distance-based indices, we construct models based on the structural properties of drug compounds. Comparative analysis with existing QSPR models employing degree and reverse degree parameters demonstrates significantly enhanced predictive capabilities of our proposed models.