芙蓉李果实成熟期间的综合品质评价指标筛选与表观预测模型构建

Selection of the comprehensive quality evaluation indexes and modeling of appearance prediction during Furongli plum ripening

中文关键词:  芙蓉李  成熟  品质指标  绝对因子分析-多元线性回归分析  表观预测模型

英文关键词:Furongli  ripening  quality indexes  absolute principal component scores-multiple linear regression  appearance prediction model

中文摘要:

      目的 科学评价芙蓉李果实成熟期间的营养品质, 建立色度值表观特征与营养品质的关系。方法 以福建省主栽品种芙蓉李为研究对象, 对其成熟期间果糖、葡萄糖、蔗糖、苹果酸、奎尼酸、琥珀酸、柠檬酸、富马酸、矢车菊素-3-芸香糖苷、矢车菊素-3-葡萄糖苷、多酚、黄酮、类胡萝卜素等13个品质指标进行分析和综合评价。结果 芙蓉李成熟期间, 各品质指标的含量变化存在显著差异(P<0.05), 综合运用相关分析、因子分析、绝对因子分析-多元线性回归(absolute principal component scores-multiple linear regression, APCS-MLR)分析筛选可反映芙蓉李综合品质的主要指标。因子分析提取出3个主因子, 贡献率分别为52.677%、23.468%、11.649%, 累计贡献率为87.794%。综合APCS-MLR等数理统计分析, 主因子1主要对果糖、矢车菊素-3-芸香糖苷、矢车菊素-3-葡萄糖苷贡献较大, 贡献率分别为53.00%、73.85%、55.54%; 主因子2主要对蔗糖、富马酸、果糖、柠檬酸的贡献率较大, 分别为28.26%、18.70%、16.14%、15.59%; 主因子3主要对多酚(29.13%)和黄酮(28.28%)有较大贡献率; 选取3个主因子总贡献率高于60%的果糖、葡萄糖、矢车菊素-3-芸香糖苷、矢车菊素-3-葡萄糖苷作为综合品质评价的主要指标。分别对已筛选出的4个主要评价指标与色度值进行多元线性逐步回归分析, 建立4个主要指标与色度值的表观预测模型, 各模型均具有较好的拟合度, 预测值与实测值的均方根误差较小; 进一步验证结果表明, 通过色度值对4个指标的预测具有较高的可靠性和准确性。结论 本研究筛选出的主要指标及预测模型可更加简单、便捷地评价芙蓉李果实成熟期间的综合品质。

英文摘要:

      Objective To scientifically evaluate the nutritional quality of Furongli plum during the ripening period, and establish the relationship between apparent characteristics of chromaticity values and the nutritional quality. Methods Furongli plum, the main cultivate in Fujian Province, was taken as the research object, and 13 quality indexes including fructose, glucose, sucrose, malic acid, quinic acid, succinic acid, citric acid, fumaric acid, cyanidin-3-rutinoside, cyanidin-3-glucoside, polyphenols, flavonoids and carotenoids were analyzed and comprehensively evaluated. Results During the ripening period, there were significant differences in the content changes of each quality index (P<0.05). Correlation analysis, factor analysis and absolute principal component scores-multiple linear regression (APCS-MLR) were comprehensively used to analyze and screen the main indexes that could reflect the comprehensive quality of Furongli plum. Factor analysis extracted three main factors, the contribution rate was 52.677%, 23.468%, 11.649%, and the cumulative contribution rate was 87.794%. APCS-MLR and other mathematical statistical analysis showed that main factor 1 mainly contributed to fructose, cyanidin-3-rutinoside and cyanidin-3-glucoside, with contribution rates of 53.00%, 73.85% and 55.54%, respectively. The contribution rate of main factor 2 to sucrose, fumaric acid, fructose and citric acid was 28.26%, 18.70%, 16.14% and 15.59% respectively. The main factor 3 mainly contributed to polyphenols (29.13%) and flavonoids (28.28%). Fructose, glucose, cyanidin-3-rutinoside and cyanidin-3-glucoside with the total contribution rate of 3 main factors higher than 60% were selected as the main indexes for comprehensive quality evaluation. Multiple linear stepwise regression analysis was carried out on the 4 main evaluation indexes and colorimetric values respectively, and the appearance prediction models of the 4 main indexes with colorimetric values were established. Each model had good fit, and the root-mean-square error between the predicted values and the measured values was small. The further verification results showed that the prediction of the 4 indexes by colorimetric value had high reliability and accuracy. Conclusion The main indexes and prediction models selected in this study can be used to evaluate the comprehensive quality of Furongli plum fruit during ripening more simply and conveniently.