Multi-Target QSAR Modelling against CYP1B1, CDK4 and Tubulin for Anticancer Drug Design

Cancer is a difficult disease to treat because tumour cells growth can depend on more than one biological pathway. Therefore, multi-target drug design may be useful for developing better anti-cancer candidates. In this study, a QSAR-based computational approach is proposed to identify potential compounds targeting CYP1B1, CDK4, and tubulin. These targets are related to tumour metabolism, cell-cycle regulation and microtubule functions. Separate datasets were obtained from BindingDB. Compounds having only direct target-based activity (IC₅₀) were taken and duplicate compounds were checked. IC₅₀ values were converted to pIC₅₀. RDkit 2D descriptors, Morgan fingerprint descriptors, and pharmacophoric features were generated in Google Colab. Three separate QSAR models are developed and validated for each target. After proper validation, the three models will be used to predict activity of our designed molecule library against all three targets. In future, the best selected candidates will be furthur evaluated using structure-based methods. Molecular docking will be performed to study their probable binding mode inside the active sited of CYP1B1, CDK4, and tubulin. The most promising docked complexes wiil be evaluated for their stability by using molecular dynamics. And finally binding free energy analysis will be performed to estimate their probable binding affinity towrds the three target proteins. The developed models will be made publicly available on GitHub to support screening of new compounds by other researchers.