中文摘要：本发明公开一种转基因豆奶粉的快速无损鉴别方法，包括步骤：(1)制备转基因和非专基因豆奶粉的测试样本，利用红外光谱仪进行红外光谱数据的采集。(2)对采集的红外光谱数据采用Savitzky?Golay平滑预处理，进行主成分分析；(3)根据预处理后的红外吸收光谱，利用PCA?loading的方法提取特征波长；(4)分别建立红外光谱和特征波长的判别分析模型，包括PLS?DA判别分析模型和/或SVM判别分析模型；(5)对于待测转基因豆奶粉和非转基因样本，依次通过步骤(1)?步骤(4)，将待测样本的红外光谱，利用所述的判别分析模型来估测类别。本发明相对于传统的分子检测方法来说，无需复杂的样本预处理，而且快速、无损、精度高。
外文摘要：NOVELTY - Performing rapid and non-destructive identification of a transgenic soybean milk powder, involves preparing transgenic and non-transgenic soybean milk powder test samples, using infrared spectroscopy to collect the infrared spectrum data; using Savitzky-Golay smoothing filter for smoothing infrared spectrum data for performing principal component analysis, using principal component analysis (PCA)-loading method to extract the characteristic wavelength, then establishing the discriminant analysis model of infrared spectrum and characteristic wavelength.
USE - Method for performing rapid and non-destructive identification of transgenic soybean milk powder (claimed).
ADVANTAGE - The method enables to perform rapid and non-destructive identification of transgenic soybean milk powder; with high precision; and does not need complicated sample pre-treatment process.
DETAILED DESCRIPTION - Performing rapid and non-destructive identification of a transgenic soybean milk powder, involves preparing transgenic and non-transgenic soybean milk powder test samples, using infrared spectroscopy to collect the infrared spectrum data; using Savitzky-Golay smoothing filter for smoothing infrared spectrum data for performing principal component analysis, using principal component analysis (PCA)-loading method to extract the characteristic wavelength, establishing the discriminant analysis model of infrared spectrum and characteristic wavelength, including partial least squares discriminant analysis (PLS-DA) model and/or support vector machine (SVM) discriminant analysis model; taking transgenic and non-transgenic soybean milk powder sample, performing infrared spectroscopy, then establishing the discriminant analysis model for estimating the category of soybean.
TF   TECHNOLOGY FOCUS - BIOLOGY - Preferred Components: The test sample comprises 50% transgenic soybean or non-transgenic soybean powder, 10% maltose, 5% edible salt, 13 freeze-dried royal jelly powder%, 7% sugar, 5% flavor, and 10% additive.
主权项：一种转基因豆奶粉的快速无损鉴别方法，其特征在于，包括步骤：(1)制备转基因和非专基因豆奶粉的测试样本，利用红外光谱仪进行红外光谱数据的采集。(2)对采集的红外光谱数据采用Savitzky?Golay平滑预处理，进行主成分分析；(3)根据预处理后的红外吸收光谱，利用PCA?loading的方法提取特征波长；(4)分别建立红外光谱和特征波长的判别分析模型，包括PLS?DA判别分析模型和/或SVM判别分析模型；(5)对于待测转基因豆奶粉和非转基因样本，依次通过步骤(1)?步骤(4)，将待测样本的红外光谱，利用所述的判别分析模型来估测类别。
申请号：CN201610931644.1
专利号：CN106338488 -A
公开/公告号：CN106338488 -A
申请/专利权人：浙江大学
发明/设计人：冯旭萍;何勇
分类号：G01N21/3563
主分类号：G01N21/3563
法律状态：实质审查
点击下载：一种转基因豆奶粉的快速无损鉴别方法