机器学习
虚拟筛选
对接(动物)
托法替尼
随机森林
人工智能
分子描述符
计算机科学
药物重新定位
支持向量机
药物发现
IC50型
化学信息学
数量结构-活动关系
药品
计算生物学
化学
药理学
生物
生物化学
医学
计算化学
体外
类风湿性关节炎
护理部
免疫学
作者
Muhammad Yasir,Jinyoung Park,Eun‐Taek Han,Won Sun Park,Jin‐Hee Han,Yong-Soo Kwon,Hee Jae Lee,Wanjoo Chun
标识
DOI:10.1021/acs.jcim.3c01090
摘要
Machine learning algorithms have been increasingly applied in drug development due to their efficiency and effectiveness. Machine learning-based drug repurposing can contribute to the identification of novel therapeutic applications for drugs with other indications. The current study used a trained machine learning model to screen a vast chemical library for new JAK2 inhibitors, the biological activities of which were reported. Reference JAK2 inhibitors, comprising 1911 compounds, have experimentally determined IC50 values. To generate the input to the machine learning model, reference compounds were subjected to RDKit, a cheminformatic toolkit, to extract molecular descriptors. A Random Forest Regression model from the Scikit-learn machine learning library was applied to obtain a predictive regression model and to analyze each molecular descriptor's role in determining IC50 values in the reference data set. Then, IC50 values of the library compounds, comprised of 1,576,903 compounds, were predicted using the generated regression model. Interestingly, some compounds that exhibit high IC50 values from the prediction were reported to possess JAK inhibition activity, which indicates the limitations of the prediction model. To confirm the JAK2 inhibition activity of predicted compounds, molecular docking and molecular dynamics simulation were carried out with the JAK inhibitor reference compound, tofacitinib. The binding affinity of docked compounds in the active region of JAK2 was also analyzed by the gmxMMPBSA approach. Furthermore, experimental validation confirmed the results from the computational analysis. Results showed highly comparable outcomes concerning tofacitinib. Conclusively, the machine learning model can efficiently improve the virtual screening of drugs and drug development.
科研通智能强力驱动
Strongly Powered by AbleSci AI