drGAT: Attention-Guided Gene Assessment of Drug Response Utilizing a Drug-Cell-Gene Heterogeneous Network

2024-05-14

arXiv.org

AI Summary

Drug response prediction suffers from limited interpretability of model outputs. To address this, we propose drGAT, the first interpretable heterogeneous graph neural network that jointly models drug–cell line–gene interactions while explicitly incorporating biological mechanisms. drGAT constructs a heterogeneous graph using Graph Attention Networks (GATs) and employs attention weights to quantify gene-level importance, thereby enabling mechanistic interpretation. These attention-derived gene priorities are further validated against PubMed literature to confirm known drug targets and regulatory relationships. Evaluated on the NCI60 benchmark, drGAT achieves 78% accuracy and precision, with an F1-score of 76%. It successfully recapitulates established top-ranked targets (e.g., TOP1) and identifies multiple novel candidate regulatory genes supported by prior literature. This work establishes a new paradigm for drug response prediction that simultaneously delivers high predictive performance and strong biological interpretability.

Technology Category

• Machine Learning: Transparent, Interpretable, Explainable ML
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Drug development is a lengthy process with a high failure rate. Increasingly, machine learning is utilized to facilitate the drug development processes. These models aim to enhance our understanding of drug characteristics, including their activity in biological contexts. However, a major challenge in drug response (DR) prediction is model interpretability as it aids in the validation of findings. This is important in biomedicine, where models need to be understandable in comparison with established knowledge of drug interactions with proteins. drGAT, a graph deep learning model, leverages a heterogeneous graph composed of relationships between proteins, cell lines, and drugs. drGAT is designed with two objectives: DR prediction as a binary sensitivity prediction and elucidation of drug mechanism from attention coefficients. drGAT has demonstrated superior performance over existing models, achieving 78% accuracy (and precision), and 76% F1 score for 269 DNA-damaging compounds of the NCI60 drug response dataset. To assess the model’s interpretability, we conducted a review of drug-gene co-occurrences in Pubmed abstracts in comparison to the top 5 genes with the highest attention coefficients for each drug. We also examined whether known relationships were retained in the model by inspecting the neighborhoods of topoisomerase-related drugs. For example, our model retained TOP1 as a highly weighted predictive feature for irinotecan and topotecan, in addition to other genes that could potentially be regulators of the drugs. Our method can be used to accurately predict sensitivity to drugs and may be useful in the identification of biomarkers relating to the treatment of cancer patients.

Problem

• Predicting drug response sensitivity using heterogeneous network
• Identifying biomarkers through attention-guided gene assessment
• Enhancing interpretability of drug-gene associations and biological processes

Innovation

• Graph deep learning model for drug response prediction
• Heterogeneous network integrating drugs, genes, cell lines
• Attention coefficients enable biomarker identification and interpretability