From Sequences to Graphs: A Graph Neural Network Framework with Attention Readout for Drug-Target Interaction Prediction

Guddu Kumar; Mr. Chhattarpal

From Sequences to Graphs: A Graph Neural Network Framework with Attention Readout for Drug-Target Interaction Prediction

Author(s)	Mr. Guddu Kumar, Mr. Chhattarpal
Country	India
Abstract	Drug-target interaction (DTI) prediction sits at the very heart of modern drug discovery. Identifying which molecular compounds bind effectively to which protein targets is a problem that has traditionally demanded enormous time, financial resources, and years of experimental validation. In this paper, we present a graph neural network-based framework that treats both drug molecules and protein structures as graphs, allowing the model to learn rich, context-aware representations directly from their topology. Our approach combines Graph Isomorphism Network (GIN) layers with attention-driven readout mechanisms, trained on three widely used benchmark datasets — BindingDB, Davis, and KIBA. Experimental results demonstrate that our method achieves an AUROC of 0.921 on BindingDB, outperforming state-of-the-art baselines including GraphDTA, DeepDTA, and AttentionDTI by a meaningful margin. Ablation studies confirm the individual contribution of each architectural component. Beyond benchmarks, we discuss the practical relevance of graphstructured molecular representations, the limitations of current formulations, and the road ahead toward three-dimensional geometric deep learning for molecular binding.
Keywords	Graph Neural Networks, Drug-Target Interaction, Deep Learning, Bioinformatics, Drug Discovery, GIN, Attention Mechanism, Molecular Graphs, Protein Contact Graphs, Computational Pharmacology.
Field	Engineering
Published In	Volume 8, Issue 3, May-June 2026
Published On	2026-05-09

View / Download PDF File

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

From Sequences to Graphs: A Graph Neural Network Framework with Attention Readout for Drug-Target Interaction Prediction

Share this