Send to

Choose Destination
Sci Rep. 2016 Nov 2;6:36325. doi: 10.1038/srep36325.

Large-scale identification of adverse drug reaction-related proteins through a random walk model.

Author information

College of Bioinformatics Science and Technology, Harbin Medical University, Harbin, 150081, China.
Center for Endemic Disease Control, Chinese Center for Disease Control and Prevention, Harbin Medical University, Harbin, 150081, China.


Adverse drug reactions (ADRs) are responsible for drug failure in clinical trials and affect life quality of patients. The identification of ADRs during the early phases of drug development is an important task. Therefore, predicting potential protein targets eliciting ADRs is essential for understanding the pathogenesis of ADRs. In this study, we proposed a computational algorithm,Integrated Network for Protein-ADR relations (INPADR), to infer potential protein-ADR relations based on an integrated network. First, the integrated network was constructed by connecting the protein-protein interaction network and the ADR similarity network using known protein-ADR relations. Then, candidate protein-ADR relations were further prioritized by performing a random walk with restart on this integrated network. Leave-one-out cross validation was used to evaluate the ability of the INPADR. An AUC of 0.8486 was obtained, which was a significant improvement compared to previous methods. We also applied the INPADR to two ADRs to evaluate its accuracy. The results suggested that the INPADR is capable of finding novel protein-ADR relations. This study provides new insight to our understanding of ADRs. The predicted ADR-related proteins will provide a reference for preclinical safety pharmacology studies and facilitate the identification of ADRs during the early phases of drug development.

[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Nature Publishing Group Icon for PubMed Central
Loading ...
Support Center