三维全息原子场作用矢量用于苦味二肽QSAR研究  

QSAR studies of bitter tasting thresholds by 3D-HoVAIF

作　　者：仝建波[1,2] 李康楠 吴英纪 占培 

机构地区：[1]陕西科技大学化学与化工学院,西安710021 [2]教育部轻化工助剂化学与技术重点实验室,西安710021

出　　处：《原子与分子物理学报》2017年第1期155-160,共6页Journal of Atomic and Molecular Physics

基　　金：国家自然科学基金(21475081);陕西省自然科学基础研究计划(2015JM2057);陕西科技大学研究生创新基金

摘　　要：本文利用三维全息原子场作用矢量(3D-Ho VAIF)对苦味二肽分子结构进行表征,并利用逐步回归结合多元线性回归建立苦味二肽定量构效关系模型,同时采用内外部双重验证的方法检验模型的稳定性.所建模型相关统计参量如下:复相关系数(R_(cum)~2)、留一法(LOO)交互验证相关系数(R_(cv)~2)、外部样本验证相关系数(Q_(ext)~2)和均方根误差(RMSE)分别为0.983、0.934、0.876和0.145.结果表明,3D-VAIF能较好地表征苦味二肽结构,优于传统的氨基酸描述子.为强活性肽类药物分子设计和改造提供指导.By applying the three -dimensional holographic vector of atomic interaction field （3D -HoVAIF） to express the structure of bitter tasting threshold of dipeptide, the quantitative structure activity relationship （QSAR） model was built by the multiple linear regression （MLR）. The obtained MLR models with the squared correlation coefficient between the predicted and observed activities （R^2cum ）, the squared cross -validation correlation coefficient （ 2 Roy ）, the external correlation coefficient （ Q^2ext ） and root - mean - square error were 0. 987, 0. 934, 0. 876 and 0. 145, respectively. The results show that 3D -HoVAIF is superior to the traditional amino acid descriptor for bitter tasting threshold of dipeptides due to the high predictive ability. Thus, it can provide guidance for the new active peptide drug molecular design and modification.

关 键 词：三维全息原子场作用矢量 苦味二肽 多元线性回归 定量构效关系 

分 类 号：O656[理学—分析化学]