Relation path feature embedding based convolutional neural network method for drug discovery

Di Zhao,Chunmei Yang,Hongfei Lin,Jiabin Wen,Jian Wang,Shengtian Sang

doi:10.1186/s12911-019-0764-5

Abstract

BackgroundDrug development is an expensive and time-consuming process. Literature-based discovery has played a critical role in drug development and may be a supplementary method to help scientists speed up the discovery of drugs.MethodsHere, we propose a relation path features embedding based convolutional neural network model with attention mechanism for drug discovery from literature, which we denote as PACNN. First, we use predications from biomedical abstracts to construct a biomedical knowledge graph, and then apply a path ranking algorithm to extract drug-disease relation path features on the biomedical knowledge graph. After that, we use these drug-disease relation features to train a convolutional neural network model which combined with the attention mechanism. Finally, we employ the trained models to mine drugs for treating diseases.ResultsThe experiment shows that the proposed model achieved promising results, comparing to several random walk algorithms.ConclusionsIn this paper, we propose a relation path features embedding based convolutional neural network with attention mechanism for discovering potential drugs from literature. Our method could be an auxiliary method for drug discovery, which can speed up the discovery of new drugs for the incurable diseases.

Highlights

Drug development is an expensive and time-consuming process
We propose a convolutional neural network (CNN) model with attention mechanism method that exploits the drug-disease relation path features for drug discovery
Results we first introduce the details of the knowledge graph (KG) and the training data, followed by several metrics used to measure the performance of our method

Summary

Introduction

Literature-based discovery has played a critical role in drug development and may be a supplementary method to help scientists speed up the discovery of drugs. Literaturebased Discovery (LBD) is a safe and low-cost technique that links the existing knowledge reported in unrelated literature sources for discovering new relationships [3, 4]. Hristovski et al introduced a semantic pattern-based LBD method which may be used to find more complex hidden associations from literature [10]. The above method ignores the relation path features information which plays an important role in LBD. Despite these considerable advances, there is still a significant room for improvement in mining drug therapies from literature

Methods

Results

Conclusion