An official website of the United States government
Drug repositioning has drawn significant attention for drug development in pharmaceutical research and industry, because of its advantages in cost and time compared with the de novo drug development. The availability of biomedical databases and online health-related information, as well as the high-performance computing, empowers the development of computational drug repositioning methods. In this work, we developed a systematic approach that identifies repositioning drugs based on heterogeneous network mining using both pharmaceutical databases (PharmGKB and SIDER) and online health community (MedHelp). By utilizing adverse drug reactions (ADRs) as the intermediate, we constructed a heterogeneous health network containing drugs, diseases, and ADRs, and developed path-based heterogeneous network mining approaches for drug repositioning. Additionally, we investigated on how the data sources affect the performance on drug repositioning. Experiment results showed that combining both PharmKGB and MedHelp identified 479 repositioning drugs, which are more than the repositioning drugs discovered by other alternatives. In addition, 31% of the 479 of the discovered repositioning drugs were supported by evidence from PubMed.
more »
« less
This content will become publicly available on August 31, 2026
Dual-route embedding-aware graph neural networks for drug repositioning
Drug repositioning presents a compelling strategy to accelerate therapeutic development by uncovering new indications for existing compounds. However, current computational methods are often limited in their ability to integrate heterogeneous biomedical data and model the intricate, multiscale relationships underlying drug–disease associations, while large-scale experimental validation remains prohibitively resource-intensive. Here, we present DREAM-GNN (Dual-Route Embedding-Aware Model for Graph Neural Networks), a multiview deep graph learning framework that incorporates biomedical domain knowledge with two complementary graphs capturing both topological structure and feature similarity to enable accurate and biologically meaningful prediction of drug–disease associations. Extensive experiments on benchmark datasets demonstrate that DREAM-GNN significantly outperforms current state-of-the-art methods in recovering artificially removed repositioning candidates, including in scenarios involving unseen drugs and diseases. These results establish DREAM-GNN as a robust and generalizable computational framework with broad potential to streamline drug discovery and advance precision medicine.
more »
« less
- Award ID(s):
- 2246157
- PAR ID:
- 10677443
- Publisher / Repository:
- Oxford Academic
- Date Published:
- Journal Name:
- Briefings in Bioinformatics
- Volume:
- 26
- Issue:
- 5
- ISSN:
- 1467-5463
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Tremendous recent literature show that associations between different brain regions, i.e., brain connectivity, provide early symptoms of neurological disorders. Despite significant efforts made for graph neural network (GNN) techniques, their focus on graph nodes makes the state-of-the-art GNN methods not suitable for classifying brain connectivity as graphs where the objective is to characterize disease-relevant network dysfunction patterns on graph links. To address this issue, we propose Multi-resolution Edge Network (MENET) to detect disease-specific connectomic benchmarks with high discrimination power across diagnostic categories. The core of MENET is a novel graph edge-wise transform that we propose, which allows us to capture multi-resolution “connectomic” features. Using a rich set of the connectomic features, we devise a graph learning framework to jointly select discriminative edges and assign diagnostic labels for graphs. Experiments on two real datasets show that MENET accurately predicts diagnostic labels and identify brain connectivities highly associated with neurological disorders such as Alzheimer’s Disease and Attention-Deficit/Hyperactivity Disorder.more » « less
-
Biomedical knowledge graphs (KGs) encode rich, structured information critical for drug discovery tasks, but extracting meaningful insights from large-scale KGs remains challenging due to their complex structure. Existing biomedical subgraph retrieval methods are tailored for graph neural networks (GNNs), limiting compatibility with other paradigms, including large language models (LLMs). We introduce K-Paths, a model-agnostic retrieval framework that extracts structured, diverse, and biologically meaningful multi-hop paths from dense biomedical KGs. These paths enable prediction of unobserved drug-drug and drug-disease interactions, including those involving entities not seen during training, thus supporting inductive reasoning. K-Paths is training-free and employs a diversity-aware adaptation of Yen's algorithm to extract the K shortest loopless paths between entities in a query, prioritizing biologically relevant and relationally diverse connections. These paths serve as concise, interpretable reasoning chains that can be directly integrated with LLMs or GNNs to improve generalization, accuracy, and enable explainable inference. Experiments on benchmark datasets show that K-Paths improves zero-shot reasoning across state-of-the-art LLMs. For instance, Tx-Gemma 27B improves by 19.8 and 4.0 F1 points on interaction severity prediction and drug repurposing tasks, respectively. Llama 70B achieves gains of 8.5 and 6.2 points on the same tasks. K-Paths also boosts the training efficiency of EmerGNN, a state-of-the-art GNN, by reducing the KG size by 90% while maintaining predictive performance. Beyond efficiency, K-Paths bridges the gap between KGs and LLMs, enabling scalable and explainable LLM-augmented scientific discovery. We release our code and the retrieved paths as a benchmark for inductive reasoning.more » « less
-
null (Ed.)Abstract Molecular interaction networks are powerful resources for molecular discovery. They are increasingly used with machine learning methods to predict biologically meaningful interactions. While deep learning on graphs has dramatically advanced the prediction prowess, current graph neural network (GNN) methods are mainly optimized for prediction on the basis of direct similarity between interacting nodes. In biological networks, however, similarity between nodes that do not directly interact has proved incredibly useful in the last decade across a variety of interaction networks. Here, we present SkipGNN, a graph neural network approach for the prediction of molecular interactions. SkipGNN predicts molecular interactions by not only aggregating information from direct interactions but also from second-order interactions, which we call skip similarity. In contrast to existing GNNs, SkipGNN receives neural messages from two-hop neighbors as well as immediate neighbors in the interaction network and non-linearly transforms the messages to obtain useful information for prediction. To inject skip similarity into a GNN, we construct a modified version of the original network, called the skip graph. We then develop an iterative fusion scheme that optimizes a GNN using both the skip graph and the original graph. Experiments on four interaction networks, including drug–drug, drug–target, protein–protein, and gene–disease interactions, show that SkipGNN achieves superior and robust performance. Furthermore, we show that unlike popular GNNs, SkipGNN learns biologically meaningful embeddings and performs especially well on noisy, incomplete interaction networks.more » « less
-
null (Ed.)Abstract The network-based proximity between drug targets and disease genes can provide novel insights regarding the repercussions, interplay, and repositioning of drugs in the context of disease. Current understanding and treatment for reversing of the fibrotic process is limited in systemic sclerosis (SSc). We have developed a network-based analysis for drug effects that takes into account the human interactome network, proximity measures between drug targets and disease-associated genes, genome-wide gene expression and disease modules that emerge through pertinent analysis. Currently used and potential drugs showed a wide variation in proximity to SSc-associated genes and distinctive proximity to the SSc-relevant pathways, depending on their class and targets. Tyrosine kinase inhibitors (TyKIs) approach disease gene through multiple pathways, including both inflammatory and fibrosing processes. The SSc disease module includes the emerging molecular targets and is in better accord with the current knowledge of the pathophysiology of the disease. In the disease-module network, the greatest perturbing activity was shown by nintedanib, followed by imatinib, dasatinib, and acetylcysteine. Suppression of the SSc-relevant pathways and alleviation of the skin fibrosis was remarkable in the inflammatory subsets of the SSc patients receiving TyKI therapy. Our results show that network-based drug-disease proximity offers a novel perspective into a drug’s therapeutic effect in the SSc disease module. This could be applied to drug combinations or drug repositioning, and be helpful guiding clinical trial design and subgroup analysis.more » « less
