Application of chemoinformatics and molecular simulations in lead optimization targeting A549 cell proliferation for lung cancer therapy

This study reports the development and validation of a statistically robust Quantitative Structure–Activity Relationship (QSAR) model for predicting the antiproliferative activity of small-molecule compounds against the A549 human lung carcinoma cell line. The work outlines a systematic approach for constructing and evaluating a predictive QSAR framework that identifies key structural determinants governing cytotoxic efficacy. A curated dataset underwent rigorous preprocessing to eliminate redundant entries, salts, and non-human bioassay data, followed by conversion of IC₅₀ values to pIC₅₀ to ensure data uniformity. Molecular descriptors were computed using PyDescriptor and subsequently refined via both objective and subjective feature selection protocols implemented in QSARINS 2. 2. 4, resulting in the identification of eight optimal descriptors contributing to model performance. Among these, the most significant; comₛpChyd₆A, comChyd₉A, fOringN3B, and nₛp3C₂B exhibited strong positive correlations with biological activity. These descriptors indicate that sp-hybridized hydrophobic carbon atoms near the molecular center of mass, increased overall hydrophobicity, and appropriately positioned nitrogen atoms enhance membrane permeability and receptor-binding affinity. In contrast, descriptors such as fNH₂B, fsp₂CnotringO₁B, and fspCC₅B were negatively correlated with activity, likely due to steric hindrance, diminished lipophilicity, and suboptimal electronic configurations. Mechanistic validation through matched molecular pair analysis confirmed the interpretability and chemical relevance of the selected descriptors, reinforcing the model’s internal consistency within its defined applicability domain. Residual diagnostics, along with Williams and Insubria plots, further validated the model’s statistical integrity, revealing minimal overfitting and a well-constrained applicability boundary. Collectively, these findings underscore the reliability and translational potential of the QSAR model as a rational design tool to guide future development of potent A549 inhibitors by emphasizing favorable structural motifs and excluding deleterious molecular features.