Artificial Intelligence Approach to Investigate the Longevity Drug

Jun-Yan Li,Hsin-Yi Chen,Wen-jie Dai,Qu-Jie Lv,Calvin Yu-Chian Chen
DOI: https://doi.org/10.1021/acs.jpclett.9b02220
2019-01-01
Abstract:Longevity is a very important and interesting topic, and Klotho has been demonstrated to be related to longevity. We combined network pharmacology, machine learning, deep learning, and molecular dynamics (MD) simulation to investigate potent lead drugs. Related protein insulin-like growth factor 1 receptor (IGF1R) and insulin receptor (IR) were docked with the traditional Chinese medicine (TCM) database to screen out several novel candidates. Besides, nine different machine learning algorithms were performed to build reliable and accurate predicted models. Moreover, we used the novel deep learning algorithm to build predicted models. All of these models obtained significant R-2, which are all greater than 0.87 on the training set and higher than 0.88 for the test set, respectively. The long time 500 ns molecular dynamics simulation was also performed to verify protein-ligand properties and stability. Finally, we obtained Antifebrile Dichroa, Holarrhena antidysenterica, and Gelsemium sempervirens, which might be potent TCMs for two targets.
What problem does this paper attempt to address?