食药物质定量结构-活性关系肺毒性预测与评估方法研究——以甘草为例OA
Prediction and Evaluation Method of Pulmonary Toxicity in Food-Medicine Substances by Quantitative Structure-Activity Relationship——Taking Glycyrrhiza uralensis as an example
目的 构建一种针对食药物质的定量结构-活性关系(QSAR)肺毒性预测与评估方法,并用于甘草的肺毒性预测.方法 在SIDER数据库中收集864种关于45类肺不良反应的化学结构,在TCMSP数据库中收集甘草所含87种化合物的分子图片,通过Stone MIND Collector软件筛选转换成SMILES形式,最后通过DrugFlow人工智能辅助药物设计平台建立QSAR肺毒性模型,采用五倍交叉验证法对已建立的模型进行准确性和稳健性的验证;TCMSP数据库和药智网数据库搜集并整合筛选得到87种甘草所含化合物,采用上述QSAR肺毒性模型对甘草进行肺毒性预测和评价.结果 对构建的QSAR肺毒性模型拟合评价,结果准确率(ACC)为0.910;平衡灵敏度与特异性-曲线下面积(ROC-AUC)为0.816,接近1,表示该模型整体预测正确率较高;预测结果显示甘草Pmax=0.436,属于"较小可能引起肺毒性"的物质,甘草引起肺毒性最有可能与化合物(2R)-2-[(E)-1-丁烯基]-5-(甲基乙氧基)-7H-异黄酮(C22H32O3),(2'R,3'S)-2',2'-二甲基-7,3',5'-三羟基-2',3'-二氢黄酮-4-内酯(C24H24O6),(3β)-羊毛甾-8,24-二烯-3-醇(C30H50O)有关.结论 本文所建立的QSAR模型能较准确地预测出食药物质潜在肺毒性强弱,并预测出可能与肺毒性相关的所含化合物,可为食药物质肺毒性的早期识别和预警提供依据.
Objective To construct a quantitative structure-activity relationship(QSAR)-based method for predicting and evaluating the pulmonary toxicity in food-medicine substances,and apply it to the prediction of pulmonary toxicity in Glycyrrhiza uralensis.Methods A total of 864 chemical structures related to 45 types of pulmonary adverse reactions were collected from the SIDER database,and molecular images of 87 compounds contained in Glycyrrhiza uralensis were collected from the TCMSP database.These molecular images were screened and converted into SMILES format using Stone MIND Collector software.Finally,a QSAR model for pulmonary toxicity was established by the DrugFlow artificial intelligence-aided drug design platform,and the five-fold cross-validation method was used to verify the accuracy and robustness of the established model.87 compounds contained in Glycyrrhiza uralensis were collected,by searching from the TCMSP database and Yaozhi Network database,and the QSAR model for pulmonary toxicity was applied to predict and evaluate the pulmonary toxicity of Glycyrrhiza uralensis.Results The constructed QSAR model for pulmonary toxicity was evaluated,with an accuracy(ACC)of 0.910,and the Receiver Operating Characteristic-Area Under the Curve(ROC-AUC)was 0.816(close to 1),indicating that the model had a relatively high overall prediction accuracy.The prediction results revealed that the Pmax of Glycyrrhiza uralensis was 0.436,classifying it as a substance of"low possibility of causing pulmonary toxicity".Glycyrrhiza uralensis induced pulmonary toxicity was most likely associated with the following compounds:(2R)-2-[(E)-1-butenyl]-5-(methylethoxy)-7H-isoflavone(C22H32O3),(2'R,3'S)-2',2'-dimethyl-7,3',5'-trihydroxy-2',3'-dihydroflavone-4-lactone(C24H24O6),and(3β)-lanosta-8,24-dien-3-ol(C30H50O).Conclusion The QSAR model established in this study can reasonably accurately predict the potential pulmonary toxicity of food-medicine substances and identify the compounds possibly related to pulmonary toxicity,which can provide a basis for the early identification and warning of pulmonary toxicity of food-medicine substances.
闵捷;付芳;李莹;吴地尧
南昌市人民医院,江西 南昌 330038联勤保障部队第九〇八医院,江西 南昌 330001江西中医药大学,江西 南昌 330004江西中医药大学,江西 南昌 330004
分子指纹定量结构-活性关系肺毒性食药物质药食同源
Molecular fingerprintQuantitative structure-activity relationshipPulmonary toxicityFood-medicine substanceHomology of food and medicine
《中医药信息》 2026 (3)
29-37,9
江西省自然科学基金项目(20224BAB206115)江西省中医药管理局科技计划项目(2023A0398)南昌市医疗卫生引导性科技计划项目(2023YLWS034)江西省卫健委科技计划项目(202311134)
评论