大理大学学报 ›› 2021, Vol. 6 ›› Issue (10): 6-.DOI: 10. 3969/j. issn. 2096-2266. 2021. 10. 002

• 药学 • 上一篇    下一篇

混合线性模型在黄酮类化合物QSAR研究中的应用

卢 昂,袁仕梦,张成桂,艾志琼,巫秀美   

  1. 1.大理大学公共卫生学院,云南大理 6710002.云南省昆虫生物医药研发重点实验室,云南大理 671000

  • 收稿日期:2021-03-25 修回日期:2021-06-07 出版日期:2021-10-15 发布日期:2021-12-11
  • 通讯作者: 巫秀美,讲师, E-mail:wxm6865@163.com。
  • 作者简介:卢昂,硕士研究生,主要从事医学统计学在医学研究中的应用研究。

Application of Mixed Linear Model in QSAR Study of Flavonoids

Lu Ang Yuan Shimeng Zhang Chenggui Ai Zhiqiong Wu Xiumei   

  1. 1. College of Public Health Dali University Dali Yunnan 671000 China 2. Yunnan Provincial Key Laboratory of Entomological Biopharmaceutical R&D Dali University Dali Yunnan 671000 China

  • Received:2021-03-25 Revised:2021-06-07 Online:2021-10-15 Published:2021-12-11

摘要:

目的:采用混合线性模型(MLM)构造定量结构-活性关系(QSAR)模型阐明黄酮类化合物与其抗氧化活性之间的构效关系,从而对该类先导化合物的设计提供依据。方法:从PubMed数据库中检索并收集黄酮类化合物对11-二苯基-2-三硝基苯肼(DPPH)的半抑制浓度(IC50)的文献,并根据其分子结构特征与IC50值建立数据库。通过MLM构造初始QSAR模型,应用五折交叉验证法进行模型优化,最终得到预测模型。结果:黄酮类化合物的抗氧化活性与A环的邻二酚羟基数目(X1)和B环的酚羟基数目(X2)呈正相关,与B环的间二酚羟基数目(X3)呈负相关。优化后的模型为:IC50=137.56-65.861-53.922+70.853。结论:采用MLM进行QSAR研究能较好地预测黄酮类化合物的抗氧化活性,为抗氧化药物的设计提供参考。

关键词: font-family:", ">QSARfont-family:宋体, ">;黄酮类化合物;抗氧化活性;混合线性模型;五折交叉验证

Abstract:

Objective To elucidate the structure-activity relationship between flavonoids and their antioxidant activities by constructing a quantitative structure-activity relationshipQSAR model using mixed linear modelMLM), hence to provide the basis for the design of these lead compounds. Methods Literature that reported the half maximal inhibitory concentrationIC50 of flavonoids against DPPH was retrieved and collected from the PubMed database and a database was established based on their molecular structure characteristics and IC50 values. The initial QSAR model was constructed through MLM the five-fold cross-validation method was applied to optimize the model and finally the prediction model was obtained. Results The antioxidant activity of flavonoid compounds was positively correlated with the numbers of ortho-phenolic-dihydroxyl in A ringX1 and the numbers of phenolic-hydroxyls in B ringX2), and negatively correlated with meta-phenolic-dihydroxyl in B ringX3. The optimized equation was that IC50 = 137.56-65.86X1-53.92 X2+70.85X3. Conclusion The established QSAR model by MLM could explicitly predict the antioxidant activity of flavonoids and provide reference for the design of antioxidants.

Key words:

"> QSAR">; flavonoids">; antioxidant activity">; mixed linear model">; five-fold cross-validation

中图分类号: