摘要
通过有机化合物分子非氢原子自身及非氢原子之间的关系构建了新的分子结构描述符,对脐橙果酒香气成分中的30个化合物进行了结构表征,采用多元线性回归(MLR)和逐步回归(SMR)建立了化合物结构与色谱保留时间的关系(QSRR)模型,同时采用"留一法"对所得模型稳健性及预测能力进行了评价。模型相关系数(R)为0.947,标准偏差(SD)为2.943;"留一法"交叉验证系数(Q2)为0.847,标准偏差(SDCV)为2.943。结果表明所构建的分子结构描述符具有较强的分子结构表达能力,模型具有良好的稳健性和预测能力,可以揭示分子结构的微观参数与保留时间的递变规律,对于预测食品中的挥发性化合物色谱保留时间、选择分离条件、探索色谱保留机理及对化合物定性识别具有重要意义。
Based on non-hydrogen atoms themselves and their relationships , 30 compounds in navel orange wine were characterized. A quantitative structure retention relationship (QSRR) model of these compounds was established through multiple linear regression and stepwise regression methods. The "leave-one-out" test was used to evaluate stability and predictive ability of the model. The traditional correlation coefficient (R) was 0. 947 and the standard deviation (SD) was 3. 578. The cross-validation correlation coefficient (Q^2) of leave-one-out (LOO) was 0. 847 and the standard deviation (SDcv) was 3. 578. The results showed that the molecular descriptor constructed in this paper could properly express structural characteristics of these compounds, and the model had good stability and predictability. The model could reveal the relationship between molecular structure' s microscopic parameters and the retention time, which was significant for the prediction of chromatographic retention time, the selection of separation condition, the exploration of chromatographic retention mechanism, and the qualitative identification of compounds.
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2015年第2期212-216,共5页
Food and Fermentation Industries
基金
四川省教育厅青年基金资助项目(13ZB0003)
内江师范学院重点项目(12NJZ01)
关键词
脐橙
果酒
香气成分
结构描述符
定量结构保留相关
navel orange
wine
aroma components
structure descriptors
quantitative structure and retention relationship