Prediction of Novel CXCR7 Inhibitors Using QSAR Modeling and Validation via Molecular Docking

TL;DR

This study developed a QSAR model using machine learning to predict CXCR7 inhibitors, validated the predictions with molecular docking, and identified promising compounds with strong binding affinities for cancer therapy.

Contribution

The paper introduces a novel QSAR modeling approach combined with molecular docking to efficiently identify potential CXCR7 inhibitors, addressing the scarcity of effective drugs targeting this receptor.

Findings

QSAR model achieved 85% accuracy in classifying active compounds

Identified several compounds with strong binding affinities (-12.24 kcal/mol)

Validated predicted inhibitors through detailed molecular docking analysis

Abstract

CXCR7, a G-protein-coupled chemokine receptor, has recently emerged as a key player in cancer progression, particularly in driving angiogenesis and metastasis. Despite its significance, currently, few effective inhibitors exist for targeting this receptor. In this study aimed to address this gap by developing a QSAR model to predict potential CXCR7 inhibitors, followed by validation through molecular docking. Using the Extra Trees classifier for QSAR modeling and employing a combination of physicochemical descriptors and molecular fingerprints, compounds were classified as active or inactive with a high accuracy of 0.85. The model could efficiently screen a large dataset, identifying several promising CXCR7 inhibitors. The predicted inhibitors were further validated through molecular docking studies, revealing strong binding affinities, with the best docking score of -12.24 +- 0.49 kcal/mol. Visualization of the docked structures in both 2D and 3D confirmed the interactions between the inhibitors and the CXCR7 receptor, reinforcing their potential efficacy.