摘要
目的以基因毒性数据(暴露浓度,CI)作为自变量建立模型,并利用该模型预测了吡咯里西啶类生物碱的基因毒性。方法本文采用偏最小二乘回归法对吡咯里西啶类生物碱的基因毒性进行预测研究。结果本文采用了PLS进行降维,进行8个成分的提取及建模,通过QSAR分析,得到吡咯里西啶类化合物的基因毒性预测模型如下:Log(toxi)=SAA*0.790049-SAG*0.141402-VOL*0.0830882+LogP*1.75971-RE*0.345487-POLAR*0.369259-TE*0.119995-BE*0.0987964-IAE*0.849541-EE*0.129214+CCI*0.0239072-HOF*0.097627-DIPOLE*0.0134722。结论本研究所建立的QSAR模型具有较好的毒性预测能力。
Objective To establish a model using genotoxicity data (exposure level, CI) as the independent variable and predict the genetic toxicity of pyrrolizidine alkaloids using the model. Methods The partial least squares method was adopted to conduct a predictive study on the genetic toxicity of pyrrolizidine alkaloids. Results Partial least squares method (PLS) was applied to reduce the dimension, 8 components were extracted and a model was built using the 8 components. The prediction model of genetic toxicity of pyrrolizidine alkaloids is obtained through quantitative structure activity relationships (QSAR) analysis. The model is as follows: Log(toxi)=SAA*0.790049-SAG*0.141402- VOL*0.0830882+LogP*1.75971-RE*0.345487-POLAR*0.369259-TE*0.119995-BE*0.0987964-IAE*0.849541- EE*0.129214+CCI*0.0239072-HOF*0.097627-DIPOLE*0.0134722. Conclusion The QSAR model established in this study has good prediction ability for toxicity.
作者
刘凯利
梁经纬
王珊
李佐静
闫心丽
孟繁浩
LIU Kaili;LIANG Jingwei;WANG San;LI Zuojing;YAN Xinli;MENG Fanhao(School of Pharmacy,China Medical University, Liaoning Shenyang 110122, China;School of Medical Devices, Shenyang PharmaceuticalUniversity, Liaoning Shenyang 110016, China)
出处
《中国药物警戒》
2019年第6期321-324,共4页
Chinese Journal of Pharmacovigilance
基金
国家自然科学基金项目(81573687
81303315
81274182):基于HSP90的乌头二萜生物碱构效关系与靶向毒效机制研究
药学统计学在中药代谢组学中生物标记物识别的研究
生物碱毒性作用靶点的识别及预测系统的研究
关键词
吡咯里西啶类生物碱
偏最小二乘回归法
基因毒性
构效关系
pyrrolizidine alkaloids
partial least squares method
genetic toxicity
quantitative structure activity relationships
