摘要: |
采用傅里叶变换近红外(FT-NIR)光谱技术, 以94份具有代表性的长牡蛎鲜样组织样本的近红外(NIR)数据和其对应的化学真实值数据为基础, 研究了NIR技术预测长牡蛎鲜样组织中水分、糖原、总蛋白质、总脂肪、锌、硒、牛磺酸和灰分8种成分含量的可行性, 通过偏最小二乘回归法(PLS)建立了以上8种成分的NIR分析模型, 并对模型进行了交互验证和外部验证。结果显示: 该实验所建立的长牡蛎鲜样组织肉质性状8项主要指标中水分、糖原和总蛋白质3种成分的NIR模型具有良好的准确性和预测能力, 建模集相关系数(RC)为0.9625-0.9902, 交互验证相关系数(RCV)为0.9342-0.9863, 外部验证的相关系数(REV)为0.9734-0.9915;但总脂肪、锌、硒、牛磺酸和灰分含量的建模效果不佳。实验结果表明, 可以运用NIR技术快速、大批量地分析长牡蛎鲜样组织中的水分、糖原和总蛋白质3种成分含量;分析过程无需使用化学试剂, 实验成本低, 且对环境无污染。长牡蛎NIR多种成分含量模型的建立, 对开展长牡蛎肉质分析、肉质性状选育、育种世代的鉴定及种质资源评价都有非常重要的意义。 |
关键词: 长牡蛎 近红外 鲜样组织 模型构建 |
DOI:10.11693/hyhz20141100305 |
分类号: |
基金项目:国家自然科学基金青年项目,31402298号;山东省农业良种工程项目——大宗经济贝类新品种选育及应用;黄河三角洲学者——海洋生物遗传育种岗位;浙江省重中之重学科开放基金,KF2015006号。 |
|
ESTABLISHMENT OF NEAR INFRARED MODELS OF EIGHT COMPONENTS ON FRESH TISSUE OF PACIFIC OYSTER CRASSOSTREA GIGAS |
WANG Wei-Jun1, YANG Jian-Min1, DONG Ying-Hui2, ZANG Heng-Chang3, WANG Zhong-Ping4, SUN Guo-Hua1
|
1.Shandong Provincial Key Laboratory of Marine Ecology Restoration, Shandong Marine Resource and Environment Research Institute, Yantai 264006, China;2.College of Biological and Environmental Sciences, Zhejiang Wanli University, Ningbo 315100, China;3.National Glycoengineering Research Center and School of Pharmaceutical Science, Shandong University, Jinan 250012, China;4.Kongdong Island Industrial Co., Ltd., Yantai 264000, China
|
Abstract: |
Using fourier transform near infrared (FT-NIR) spectrum technology, we established models on the prediction feasibility for moisture, glycogen, protein, total fat, zinc, selenium, taurine, and ash content in Pacific oysters Crassostrea gigas based on NIR diffuse reflective spectra and chemical measurement data of 94 fresh tissue samples by partial least square (PLS) regression.The models were then validated by cross validation and external validation.The results present that the NIR models performed well for predicting moisture, glycogen, and protein, in correlation coefficient in calibration (RC) between 0.9625-0.9902, correlation coefficient in cross validation (RCV) in 0.9342-0.9863, and correlation coefficient in external prediction (REV) in 0.9734-0.9915.However, these indicators did not work well for total fat, zinc, selenium, taurine, and ash.Our experimental data show that the NIR technique could be used to determine the moisture, glycogen, and protein in C.gigas rapidly, massively, and accurately using zero chemical reagent.Therefore, this approach is meaningful for evaluating the flesh quality timely and handy, as well as in flesh traits genetic selection and the germ plasm resource protection. |
Key words: Crassostrea gigas near infrared (NIR) spectrum fresh tissue sample modeling |