花生自然风干种子蔗糖含量近红外定量分析模型构建
2018-08-14唐月异王秀贞刘婷吴琪孙全喜王志伟张欣王传堂邵俊飞
唐月异 王秀贞 刘婷 吴琪 孙全喜 王志伟 张欣 王传堂 邵俊飞
摘要:为快速测定花生种子蔗糖含量,筛选口感好的花生品种,本研究采集167份自然风干花生种子的近红外光谱,并利用高效液相色谱(HPLC)法测定其蔗糖含量,然后利用偏最小二乘法(Modified PLS)构建多粒自然风干花生种子样品蔗糖含量的近红外定量分析模型,经内部交叉检验和优化,确定该模型的最佳光谱预处理方法为最小-最大归一化法,蔗糖含量谱区范围为4 597.7~11 988.0 cm-1,维数为10,模型的决定系数(R2)为81.59,交叉验证根均方差(RMSECV)为0.414。经外部验证,所建模型可以较好地预测花生种子蔗糖含量。
关键词:花生;自然风干种子;蔗糖含量;近红外定量分析
中图分类号:S565.201文献标识号:A文章编号:1001-4942(2018)06-0159-04
Abstract For rapidly determining sucrose content in peanut seeds and selecting peanut varieties with good taste, the near infrared spectroscopy model was developed to predict sucrose content of sun-dried peanut seeds. Near infrared spectrum of a total of 167 peanut samples were collected, and the sucrose content of each sample was determined by high performance liquid chromatography (HPLC). Then a NIR calibration model was established by partial least squares method (modified PLS). Through internal cross validation and optimization, the optimized spectral preprocessing method of this model was minimum and maximum normalization, and the spectral range of sucrose, rank, determination coefficient (R2) and root-mean-squares error of cross-validation (RMSECV) were 4 597.7~11 988.0 cm-1, 10, 81.59 and 0.414, respectively. Through external validation, the model could be used to predict the sucrose content of peanut seeds very well.
Keywords Peanut; Sun-dried seed; Sucrose content; Near infrared spectroscopy (NIRS)
花生是我國重要的经济作物,其籽仁含丰富的营养物质,其中脂肪含量约占50%,蛋白质含量约占24%~36%,碳水化合物约占10%~23%。蔗糖是碳水化合物的重要成分,也是花生甜味的主要来源,其含量的多少与焙烤花生果(仁)的香气和味道有密切关系[1]。Goch等研究表明花生仁的甜度与口味品质间的相关系数达到0.88,且当蔗糖含量超过5%时这种趋势尤为突出[2]。据报道,我国和美国出产的花生比阿根廷出产的蔗糖含量低[3]。日本市场要求花生蔗糖含量达到6%以上,而国内生产上应用的多数花生品种蔗糖含量很难达到这个要求,这成为我国花生出口日本的主要限制因素[4]。Pattee等(2000)[5]发现甜味是可以遗传的性状,培育蔗糖含量高的花生品种,不仅能提升花生的感官品质,也可以更好地满足国际市场的需求。……
