Objective Pharmacological methods were used to screen targets and signaling pathways of Ma Xing Shi Gan Decoction(MXSGD)during influenza treatments,and mechanisms underlying antiinfluenza effects were elucidated.Metho...Objective Pharmacological methods were used to screen targets and signaling pathways of Ma Xing Shi Gan Decoction(MXSGD)during influenza treatments,and mechanisms underlying antiinfluenza effects were elucidated.Methods The Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform(TCMSP)and relevant literature were searched under predefined conditions to identify the main compounds and their targets.Interactions between the target proteins were predicted using the STRING database.Gene Ontology(GO)functional enrichment analyses and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway analyses were performed on the core targets involved in the influenza protein-protein interaction(PPI)network,using WebGestalt and the reactome database.iGEMDOCK was used for molecular docking of receptors and ligands to produce docking scores,and the results were visualized using Autodock and PyMOL.Results In total,126 major compounds and their respective targets were screened.355 influenza target proteins and 1221 influenza protein interactions were predicted using the STRING database.Influenza-related signaling pathways were strongly enriched in pharmacodynamic targets of MXSGD such as cytokine signaling in immune system and signaling by interleukin.The main biological process was response to the stimulates.Molecular docking results showed that RELALicochalcone A docking elicited by MXSGD,was superior to that of other target proteins and active compounds,suggesting that the docking site is also the main effector site of MXSGD during influenza treatments.Conclusions The results showed that MXSGD exerts antiinfluenza effects by interfering with virus adsorption,inhibiting virus proliferation,influencing immune functions and protecting host cells,which may prevent inflammation-induced tissue damage.展开更多
目的:基于MRI影像组学构建机器学习模型,探讨其预测子宫内膜癌(EC)病理分级的可行性。方法:回顾性纳入2018年11月—2021年3月梅州市人民医院收治的162例EC患者(高级别组46例、低级别组116例),所有患者均行MRI检查(含T2WI、DWI、ADC序列...目的:基于MRI影像组学构建机器学习模型,探讨其预测子宫内膜癌(EC)病理分级的可行性。方法:回顾性纳入2018年11月—2021年3月梅州市人民医院收治的162例EC患者(高级别组46例、低级别组116例),所有患者均行MRI检查(含T2WI、DWI、ADC序列)。手动勾画肿瘤感兴趣容积(VOI),通过联影智能uAI Research Portal提取影像组学特征。采用单因素分析与LASSO回归筛选最优特征后,构建逻辑回归、支持向量机、高斯过程和随机森林4种机器学习模型,以7∶3比例划分训练集与测试集,采用受试者工作特征(ROC)曲线、校准曲线及决策曲线评估模型效能。结果:最终筛选出11个最优特征(含5个T2WI特征、4个DWI特征、2个ADC特征);训练集曲线下面积(AUC)分别为0.915(逻辑回归)、0.935(支持向量机)、0.937(高斯过程)、0.966(随机森林),测试集AUC分别为0.884(逻辑回归)、0.882(支持向量机)、0.878(高斯过程)、0.806(随机森林);校准曲线显示各模型预测概率与真实风险高度一致,决策曲线提示模型具有稳定临床净获益。其中,逻辑回归模型在训练集与测试集均保持较高且稳定的性能,鲁棒性及临床适用性更优;随机森林模型存在过拟合风险,泛化能力不足。结论:基于MRI影像组学构建的机器学习模型可有效预测EC病理分级,其中逻辑回归模型可作为临床优选方案。展开更多
基金We thank for the funding support from the National Natural Science Foundation of China(No.81973670)the Natural Science Foundation of Hunan Province(No.2018JJ2297)+1 种基金the Key Program of Scientific Research Fund of Hunan Provincial Education Department(No.19A370)the Project of Research Learning and Innovative Experiment for College Students in Hunan(No.2016284,No.2016281,No.2017281and No.2018420).
文摘Objective Pharmacological methods were used to screen targets and signaling pathways of Ma Xing Shi Gan Decoction(MXSGD)during influenza treatments,and mechanisms underlying antiinfluenza effects were elucidated.Methods The Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform(TCMSP)and relevant literature were searched under predefined conditions to identify the main compounds and their targets.Interactions between the target proteins were predicted using the STRING database.Gene Ontology(GO)functional enrichment analyses and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway analyses were performed on the core targets involved in the influenza protein-protein interaction(PPI)network,using WebGestalt and the reactome database.iGEMDOCK was used for molecular docking of receptors and ligands to produce docking scores,and the results were visualized using Autodock and PyMOL.Results In total,126 major compounds and their respective targets were screened.355 influenza target proteins and 1221 influenza protein interactions were predicted using the STRING database.Influenza-related signaling pathways were strongly enriched in pharmacodynamic targets of MXSGD such as cytokine signaling in immune system and signaling by interleukin.The main biological process was response to the stimulates.Molecular docking results showed that RELALicochalcone A docking elicited by MXSGD,was superior to that of other target proteins and active compounds,suggesting that the docking site is also the main effector site of MXSGD during influenza treatments.Conclusions The results showed that MXSGD exerts antiinfluenza effects by interfering with virus adsorption,inhibiting virus proliferation,influencing immune functions and protecting host cells,which may prevent inflammation-induced tissue damage.
文摘目的:基于MRI影像组学构建机器学习模型,探讨其预测子宫内膜癌(EC)病理分级的可行性。方法:回顾性纳入2018年11月—2021年3月梅州市人民医院收治的162例EC患者(高级别组46例、低级别组116例),所有患者均行MRI检查(含T2WI、DWI、ADC序列)。手动勾画肿瘤感兴趣容积(VOI),通过联影智能uAI Research Portal提取影像组学特征。采用单因素分析与LASSO回归筛选最优特征后,构建逻辑回归、支持向量机、高斯过程和随机森林4种机器学习模型,以7∶3比例划分训练集与测试集,采用受试者工作特征(ROC)曲线、校准曲线及决策曲线评估模型效能。结果:最终筛选出11个最优特征(含5个T2WI特征、4个DWI特征、2个ADC特征);训练集曲线下面积(AUC)分别为0.915(逻辑回归)、0.935(支持向量机)、0.937(高斯过程)、0.966(随机森林),测试集AUC分别为0.884(逻辑回归)、0.882(支持向量机)、0.878(高斯过程)、0.806(随机森林);校准曲线显示各模型预测概率与真实风险高度一致,决策曲线提示模型具有稳定临床净获益。其中,逻辑回归模型在训练集与测试集均保持较高且稳定的性能,鲁棒性及临床适用性更优;随机森林模型存在过拟合风险,泛化能力不足。结论:基于MRI影像组学构建的机器学习模型可有效预测EC病理分级,其中逻辑回归模型可作为临床优选方案。
