Drug Discovery Using Graph Neural Networks: A Deep Learning Framework for Molecular Property Prediction and Virtual Screening | IJCT Volume 13 – Issue 3 | IJCT-V13I3P116
International Journal of Computer Techniques
ISSN 2394-2231
Author
Sharayu N. Bonde, Yogita H. Dhande
Abstract
Drug discovery is a complex, expensive, and time-consuming process that traditionally requires extensive laboratory experimentation and high-throughput screening. Recent advances in artificial intelligence have enabled computational approaches for accelerating the identification of promising drug candidates. Graph Neural Networks (GNNs) have emerged as powerful models for molecular representation learning because chemical compounds can naturally be represented as graphs where atoms are nodes and chemical bonds are edges.
This study proposes a Graph Neural Network-based framework for molecular property prediction and virtual screening in drug discovery. The proposed system utilizes molecular graph representations extracted from public drug databases and applies Graph Convolutional Networks (GCN), Graph Attention Networks (GAT), and Message Passing Neural Networks (MPNN) to predict biological activity and drug-likeness properties. Experimental evaluation on benchmark molecular datasets demonstrates superior predictive performance compared to traditional machine learning approaches including Random Forest, Support Vector Machine, and Deep Neural Networks.
Results indicate that the proposed GNN framework achieves an accuracy of 92.8%, ROC-AUC of 0.95, and F1-score of 0.91, outperforming conventional methods by significant margins. The findings highlight the effectiveness of graph-based deep learning in accelerating early-stage drug discovery and reducing experimental costs.
Keywords
Drug Discovery, Graph Neural Networks, Deep Learning, Molecular Property Prediction, Virtual Screening, Graph Convolution Networks.
Conclusion
This research presented a Graph Neural Network-based framework for drug discovery and molecular property prediction. The proposed system leverages graph representations of molecular structures and advanced message-passing mechanisms to learn complex chemical interactions. Experimental evaluation on benchmark datasets demonstrated that the proposed GNN framework achieves an accuracy of 92.8% and ROC-AUC of 0.95, outperforming traditional machine learning and deep learning approaches. The study confirms that GNNs provide an effective solution for accelerating virtual screening and reducing drug discovery costs.
Future work will focus on integrating transformer-based graph architectures, explainable AI techniques, and generative molecular design models for next-generation drug discovery systems.
References
1.Corso, G., Stark, H., Jegelka, S., Jaakkola, T., & Barzilay, R. (2024). Graph Neural Networks. Nature Reviews Methods Primers, 4, 17.
2.Wieder, O., Kohlbacher, S., Kuenemann, M., et al. (2020). A Compact Review of Molecular Property Prediction with Graph Neural Networks. Drug Discovery Today: Technologies, 37, 1–12.
3. Bongini, P., Pancino, N., Bendjeddou, A., et al. (2024). Composite Graph Neural Networks for Molecular Property Prediction. International Journal of Molecular Sciences, 25(12), 6583.
4.Buterez, D., Janet, J.P., Kiddle, S.J., et al. (2024). Transfer Learning with Graph Neural Networks for Improved Molecular Property Prediction in the Multi-Fidelity Setting. Nature Communications, 15, 1517.
5.Ramachandran, A. (2024). Graph Neural Networks for Molecular Property Prediction in Drug Discovery: A Comprehensive Review. Research Review Article.
6.Wang, Y., et al. (2024). A Review on Graph Neural Networks for Predicting Synergistic Drug Combinations. Artificial Intelligence Review, 57, 49.
7.Tropsha, A., Isayev, O., Varnek, A., et al. (2024). Integrating QSAR Modelling and Deep Learning in Drug Discovery: The Emergence of Deep QSAR. Nature Reviews Drug Discovery, 23, 141–155.
8.Vamathevan, J., Clark, D., Czodrowski, P., et al. (2019). Applications of Machine Learning in Drug Discovery and Development. Nature Reviews Drug Discovery, 18, 463–477.
9.Hou, T., et al. (2025). Graph Neural Networks in Modern AI-Aided Drug Discovery. Chemical Reviews, 125(20), 10001–10103.
How to Cite This Paper
Sharayu N. Bonde, Yogita H. Dhande (2026). Drug Discovery Using Graph Neural Networks: A Deep Learning Framework for Molecular Property Prediction and Virtual Screening. International Journal of Computer Techniques, 13(3). ISSN: 2394-2231.
Drug Discovery Using Graph Neural Networks A Deep Learning Framework for Molecular Property Prediction and Virtual ScreeningDownload
Related Posts:
