In silico drug repositioning using deep learning and comprehensive similarity measures

Abstract

Background Drug repositioning, meanings finding new uses for existing drugs, which can accelerate the processing of new drugs research and development. Various computational methods have been presented to predict novel drug–disease associations for drug repositioning based on similarity measures among drugs and diseases. However, there are some known associations between drugs and diseases that previous studies not utilized. Methods In this work, we develop a deep gated recurrent units model to predict potential drug–disease interactions using comprehensive similarity measures and Gaussian interaction profile kernel. More specifically, the similarity measure is used to exploit discriminative feature for drugs based on their chemical fingerprints. Meanwhile, the Gaussian interactions profile kernel is employed to obtain efficient feature of diseases based on known disease-disease associations. Then, a deep gated recurrent units model is developed to predict potential drug–disease interactions. Results The performance of the proposed model is evaluated on two benchmark datasets under tenfold cross-validation. And to further verify the predictive ability, case studies for predicting new potential indications of drugs were carried out. Conclusion The experimental results proved the proposed model is a useful tool for predicting new indications for drugs or new treatments for diseases, and can accelerate drug repositioning and related drug research and discovery.

EKGDR: An End-to-End Knowledge Graph-Based Method for Computational Drug Repurposing

“…Second, drug repurposing has a shorter and more streamlined drug discovery and development process, which means significant savings in both time and cost. 6 In recent years, the development and availability of publicly accessible online databases on drug-related information has provided a valuable opportunity for the computational study and prediction of new drug-disease interactions and drug repurposing. Therefore, the use of computational models based on machine learning, especially deep learning, to leverage the information available in the aforementioned databases has become one of the most promising research topics in the field of drug discovery.…”

“…First, the safety of approved drugs has already been established. Second, drug repurposing has a shorter and more streamlined drug discovery and development process, which means significant savings in both time and cost …”

“…x z (6) where e x (l-1) , e z (l-1) , and e v (l-1) denote the representations of disease x, drug z, and entity v, respectively, and they include the propagated relational signals from their neighbors up to l − 1 hops. By employing such a relational modeling approach, these representations can capture comprehensive semantics of the multihop paths and consider the relational dependencies as well.…”

DDIT: An Online Predictor for Multiple Clinical Phenotypic Drug-Disease Associations

“…In addition to molecular docking, machine learning can also use structural data to make predictions. In this way, Hu et al used convolutional neural networks to predict drug–target interactions based on drug structure and protein sequences ( Hu et al, 2019 ); Yi et al developed a deep gated recurrent units model to predict potential drug–disease interactions using comprehensive similarity measures and Gaussian interaction profile kernel ( Yi et al, 2021 ); and Ke et al established a deep neural network (DNN) to identify potential drugs for anti-coronavirus activities ( Ke et al, 2020 ).…”

Identification of Repurposable Drugs for Colorectal Cancer Using Drug-Network-Based Classification Models

“…Numerous association data among drugs, diseases, proteins, and gene ontology (GO) were utilized in the methods of topological similarity and singular value decomposition (TS-SVD) [11] and meta-path based gene ontology profiles in order to predict drug-disease associations (MGP-DDA) [12]. Over the last few years, several methods have been developed based on various deep learning models such as a model for potential drug-disease interactions prediction (DDIPred) [13] and a method for identifying drug-disease associations based on a geometric deep learning framework (DDAGDL) [14]. Although there are many existing methods for generic drug repurposing, deploying those methods in identifying repurposable drugs for a particular disease is often difficult due to numerous training data required and challenges in implementation.…”