|
|
GNN-MTB: An Anti-Mycobacterium Drug Virtual Screening Method based on Graph Neural Network
|
Gu Yaowen,Zheng Si,Li Jiao
|
(Institute of Medical Information, Chinese Academy of Medical Sciences, Beijing 100020,China)
|
|
|
Abstract
[Objective] This study aims to construct an anti-tuberculosis drug virtual screening model for the research and development of anti-tuberculosis drugs.
[Methods] We proposed a curriculum learning-optimized graph neural network model for anti-tuberculosis inhibitors virtual screening, which called GNN-MTB. Furthermore, a benchmark dataset for anti-tuberculosis drugs was collected from the public database, then we compared the performance of GNN-MTB with four classic machine learning models and two graph neural network models on the benchmark dataset.
[Results] We collected and integrated 10,789 available anti-tuberculosis drug screening experimental data as our benchmark dataset. Our proposed GNN-MTB model achieved the area under the receiver operating characteristic curve (AUC) score of 0.912 and the area under precision-recall curve (AUPR) score of 0.679, which were higher than machine learning and graph neural network models. The maximum improvement of our proposed method in average AUC and average Precision are 3.87% and 13.17%. Moreover, we open-resourced the GNN-MTB and an anti-tuberculosis drug virtual screening tool for the use of anti-tuberculosis drug researchers (https://github.com/gu-yaowen/GNN-MTB).
[Limitations] The modeling of virtual screening model for anti-tuberculosis can be further integrated the analysis of drug sensitivity and bacterial resistance.
[Conclusions] The GNN-MTB proposed in our study achieved considerable performance in the anti-tuberculosis inhibitor predictions, which can be applied to the process of anti-tuberculosis drug screening. Furthermore, this method can also provide support for the construction of other disease/target-specific drug virtual screening models.
|
Published: 13 July 2022
|
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|