In recent years, based on our previously developed Network-Based Inference (NBI) method [1], we developed two improved methods for prediction of drug-target interactions (DTIs), namely substructure-drug-target NBI (SDTNBI) [2] and balanced SDTNBI (bSDTNBI) [3]. These two methods can predict potential targets for various types of compounds, not only for approved drugs, but also for other natural and unnatural products in research and development [2-4]. Compared with traditional methods such as structure-based methods and supervised machine learning, our NBI methods do not rely on three-dimensional structures of targets or negative samples [5]. Using bSDTNBI in combination with in vitro experiments, we identified new active compounds for several target proteins, including estrogen receptor α (ERα) [3], prostaglandin E2 receptor EP4 subtype (EP4) [6], and NAD(P)H:quinone oxidoreductase 1 (NQO1) [7].

More recently, we developed a novel method named weighted SDTNBI (wSDTNBI) [8]. Compared with previous network-based methods, wSDTNBI can output prediction scores correlated with binding affinities. Using this method in combination with in vitro experiments, we discovered several new inverse agoints for retinoid-related orphan receptor γt (RORγt). In vivo experiments further showed that two of them have therapeutic effects on an autoimmune disease named multiple sclerosis. These successful applications demonstrated the practical value of our NBI methods in target prediction and virtual screening.

References:

1. Feixiong Cheng^#, Chuang Liu^#, Jing Jiang, Weiqiang Lu, Weihua Li, Guixia Liu, Weixing Zhou*, Jin Huang*, Yun Tang*. Prediction of drug-target interactions and drug repositioning via network-based inference. PLoS Computational Biology, 2012, 8(5): e1002503. Download
2. Zengrui Wu, Feixiong Cheng*, Jie Li, Weihua Li, Guixia Liu, Yun Tang*. SDTNBI: an integrated network and chemoinformatics tool for systematic prediction of drug-target interactions and drug repositioning. Briefings in Bioinformatics, 2017, 18(2): 333-347. Download
3. Zengrui Wu^#, Weiqiang Lu^#, Dang Wu, Anqi Luo, Hanping Bian, Jie Li, Weihua Li, Guixia Liu, Jin Huang*, Feixiong Cheng*, Yun Tang*. In silico prediction of chemical mechanism of action via an improved network-based inference method. British Journal of Pharmacology, 2016, 173(23): 3372-3385. Download
4. Jiansong Fang^#, Zengrui Wu^#, Chuipu Cai, Qi Wang, Yun Tang*, Feixiong Cheng*. Quantitative and systems pharmacology. 1. In silico prediction of drug-target interactions of natural products enables new targeted cancer therapy. Journal of Chemical Information and Modeling, 2017, 57(11): 2657-2671. Download
5. Zengrui Wu, Weihua Li, Guixia Liu, Yun Tang*. Network-based methods for prediction of drug-target interactions. Frontiers in Pharmacology, 2018, 9: 1134. Download
6. Zengrui Wu^#, Weiqiang Lu^#, Weiwei Yu, Tianduanyi Wang, Weihua Li, Guixia Liu, Hankun Zhang, Xiufeng Pang, Jin Huang, Mingyao Liu*, Feixiong Cheng*, Yun Tang*. Quantitative and systems pharmacology 2. In silico polypharmacology of G protein-coupled receptor ligands via network-based approaches, Pharmacological Research, 2018, 129: 400-413. Download
7. Zengrui Wu^#, Qiaohui Wang^#, Hongbin Yang, Jiye Wang, Weihua Li, Guixia Liu, Yi Yang, Yuzheng Zhao*, Yun Tang*. Discovery of natural products targeting NQO1 via an approach combining network-based inference and identification of privileged substructures. Journal of Chemical Information and Modeling, 2021, 61(5): 2486-2498. Download
8. Zengrui Wu^#, Hui Ma^#, Zehui Liu^#, Lulu Zheng, Zhuohang Yu, Shuying Cao, Wenqing Fang, Lili Wu, Weihua Li, Guixia Liu, Jin Huang*, Yun Tang*. wSDTNBI: a novel network-based inference method for virtual screening. Chemical Science, 2022, 13(4): 1060-1079. Download

Besides target proteins, NBI methods can also be used for prediction of microRNAs [1]. In a previous study, we applied NBI to predict potential microRNAs for anticancer drugs, and validated several predicted drug-microRNA associations [2].

References:

1. Jie Li, Zengrui Wu, Feixiong Cheng, Weihua Li, Guixia Liu, Yun Tang*. Computational prediction of microRNA networks incorporating environmental toxicity and disease etiology. Scientific Reports, 2014, 4: 5576. Download
2. Jie Li, Kecheng Lei, Zengrui Wu, Weihua Li, Guixia Liu, Jianwen Liu*, Feixiong Cheng*, Yun Tang*. Network-based identification of microRNAs as potential pharmacogenomic biomarkers for anticancer drugs. Oncotarget, 2016, 7(29): 45584-45596. Download

In addition to target prediction, the applications of NBI methods can be further expanded to other types of prediction, by recognizing other prediction objects as special targets. For example, we developed a method derived from bSDTNBI to predict pathways [1].

Reference:

1. Jiye Wang^#, Zengrui Wu^#, Yayuan Peng, Weihua Li, Guixia Liu, Yun Tang*. Pathway-based drug repurposing with DPNetinfer: a method to predict drug-pathway associations via network-based approaches. Journal of Chemical Information and Modeling, 2021, 61(5): 2475-2485. Download

We developed a new method derived from SDTNBI to predict ATC codes. Several drugs predicted to have therapeutic effects on heart failure showed cardioprotective activities in in vitro experiments [1].

Reference:

1. Yayuan Peng^#, Manjiong Wang^#, Yixiang Xu^#, Zengrui Wu, Jiye Wang, Chao Zhang, Guixia Liu, Weihua Li, Jian Li*, Yun Tang*. Drug repositioning by prediction of drug's Anatomical Therapeutic Chemical code via network based inference approaches. Briefings in Bioinformatics, 2021, 22(2): 2058-2072. Download

In a previous study, we developed a method derived from NBI to predict ADEs [1]. However, this method cannot make predictions for novel compounds outside the known drug-ADE association network. Recently, based on our comprehensive database for adverse drug events [2], we developed a new method named ADENet [3]. Several predictions were validated by the literature.

References:

1. Feixiong Cheng, Weihua Li, Xichuan Wang, Yadi Zhou, Zengrui Wu, Jie Shen, Yun Tang*. Adverse drug events: database construction and in silico prediction. Journal of Chemical Information and Modeling, 2013, 53(4), 744-752. Download
2. Zhuohang Yu, Zengrui Wu*, Weihua Li, Guixia Liu, Yun Tang*. MetaADEDB 2.0: a comprehensive database on adverse drug events. Bioinformatics, 2021, 37(15): 2221-2222. Download
3. Zhuohang Yu, Zengrui Wu*, Weihua Li, Guixia Liu, Yun Tang*. ADENet: a novel network-based inference method for prediction of drug adverse events. Briefings in Bioinformatics, 2022, 23(2): bbab580. Download

Several years ago, we developed a standalone software named NetInfer via C++ programming language. The aforementioned network-based methods including NBI, SDTNBI and bSDTNBI were implemented in this program.

The first version of NetInfer is light weight and does not need the support of any third-party math libraries. Recently, we further improved it to accelerate the calculation. The NVIDIA cuBLAS library, an GPU-accelerated implementation of the basic linear algebra subroutines (BLAS), was used to accelerate the matrix computation such as matrix multiplication.

The latest version of NetInfer can be downloaded here. This binary executable file was compiled by GNU C++ compiler (version 4.8.5) with the support of NVIDIA CUDA toolkit (version 10.1), and has been tested on CentOS 7.

Although the standalone version of NetInfer provides a uniform platform for users to construct network-based models, evaluate model performance and perform prediction, there are still two limitations.

Firstly, the software is just a command-line program. No graphic user interface was provided. Users need to type a long command line with many arguments to run it. Even small mistakes in the command line could lead to unexpected errors.

Secondly, the software requires users to store their network data in a specified format before using it. Users may need to write programs for format conversion.

These limitations make the software unfriendly to users, especially novice ones. To improve user experience, we developed the online version of NetInfer as a web server based on the standalone version. This web server was implemented using Apache, MariaDB and PHP.