为满足不同种类食品对大豆分离蛋白(soybean protein isolate,SPI)不同功能性的需求,本研究利用红外光谱快速采集70组不同pH值处理后SPI的数据,探讨pH值变化对SPI结构含量的影响。使用均值中心化、多元散射校正、标准正态变量变换和归...为满足不同种类食品对大豆分离蛋白(soybean protein isolate,SPI)不同功能性的需求,本研究利用红外光谱快速采集70组不同pH值处理后SPI的数据,探讨pH值变化对SPI结构含量的影响。使用均值中心化、多元散射校正、标准正态变量变换和归一化算法对红外光谱数据进行预处理,基于二维相关红外光谱提取特征波段,再利用偏最小二乘(partial least square,PLS)法和算术优化算法-随机森林(arithmetic optimization algorithm-random forests,AOA-RF)建立不同pH值条件下SPI结构及含量的预测模型。结果表明,经均值中心化和多元散射校正结合处理后,α-螺旋、β-折叠、β-转角和无规卷曲模型的相对标准偏差分别为1.29%、1.60%、1.37%、7.28%,两者结合对光谱数据的预处理效果最佳。预测α-螺旋和β-折叠含量最优模型为AOA-RF(特征波段),校正集决定系数为0.9350和0.9266,预测集决定系数为0.8568和0.8701;预测β-转角和无规卷曲含量最优模型为PLS(特征波段),校正集决定系数为0.9154和0.8817,预测集决定系数为0.8913和0.7843。本研究结果可为工业生产过程中产品质量快速检测和工艺条件控制提供理论支撑。展开更多
Aim:The fast and nondestructive identification of Cortex Periplocae and Cortex Acanthopanacis by Fourier transform near infrared spectroscopy and generalized two-dimensional(2D) correlation spectroscopy were perform...Aim:The fast and nondestructive identification of Cortex Periplocae and Cortex Acanthopanacis by Fourier transform near infrared spectroscopy and generalized two-dimensional(2D) correlation spectroscopy were performed.Methods:Near infrared spectra of Cortex Periplocae and Cortex Acanthopanacis was used for computation of the 2D correlation spectra and comparison of the synchronous and asynchronous 2D correlation spectra.Results:Two of the analytes had similar spectral characteristics in one-dimensional near infrared while quite different in 2D correlation spectra.Cortex Periplocae and Cortex Acanthopanacis were identified visually using the 2D correlation spectra in region 5 600~4 700 cm-1and 7 200~6 600 cm-1.Conclusion:The results prove that 2D correlation spectra could enhance the resolution of near infrared spectra and increase the capacity of identification.The combination of 2D correlation spectroscopy and near infrared spectroscopy allow a new and convenient approach for evaluating the medicinal herbs.展开更多
旨在建立基于二维相关红外光谱(2D IR spectra)的分析技术,以进行中药中是否掺杂化学药物的判别。以减肥药盐酸芬氟拉明(fenfluramine hydrochloride,FH)、盐酸西布曲明(sibutramine hydrochloride,SH)为待检化合物,确定二者同步相关谱...旨在建立基于二维相关红外光谱(2D IR spectra)的分析技术,以进行中药中是否掺杂化学药物的判别。以减肥药盐酸芬氟拉明(fenfluramine hydrochloride,FH)、盐酸西布曲明(sibutramine hydrochloride,SH)为待检化合物,确定二者同步相关谱中的特征相关峰。首先通过比较掺杂中药与盐酸芬氟拉明的同步谱中特征相关峰的相似性,初步对掺杂化合物盐酸芬氟拉明进行判定,而后利用掺杂中药异步谱能确认相关峰来源的特性,进一步确证盐酸芬氟拉明的存在;此外,根据掺杂中药同步谱中未出现与盐酸西布曲明同步谱相似的特征相关峰,可以直接判别其中未掺杂盐酸西布曲明。实验所建立的方法无需样品分离,速度快,成本低,为快速简便地进行复杂体系中是否存在化学药物的判别提供了一种新的可能的方法。展开更多
文摘为满足不同种类食品对大豆分离蛋白(soybean protein isolate,SPI)不同功能性的需求,本研究利用红外光谱快速采集70组不同pH值处理后SPI的数据,探讨pH值变化对SPI结构含量的影响。使用均值中心化、多元散射校正、标准正态变量变换和归一化算法对红外光谱数据进行预处理,基于二维相关红外光谱提取特征波段,再利用偏最小二乘(partial least square,PLS)法和算术优化算法-随机森林(arithmetic optimization algorithm-random forests,AOA-RF)建立不同pH值条件下SPI结构及含量的预测模型。结果表明,经均值中心化和多元散射校正结合处理后,α-螺旋、β-折叠、β-转角和无规卷曲模型的相对标准偏差分别为1.29%、1.60%、1.37%、7.28%,两者结合对光谱数据的预处理效果最佳。预测α-螺旋和β-折叠含量最优模型为AOA-RF(特征波段),校正集决定系数为0.9350和0.9266,预测集决定系数为0.8568和0.8701;预测β-转角和无规卷曲含量最优模型为PLS(特征波段),校正集决定系数为0.9154和0.8817,预测集决定系数为0.8913和0.7843。本研究结果可为工业生产过程中产品质量快速检测和工艺条件控制提供理论支撑。
文摘Aim:The fast and nondestructive identification of Cortex Periplocae and Cortex Acanthopanacis by Fourier transform near infrared spectroscopy and generalized two-dimensional(2D) correlation spectroscopy were performed.Methods:Near infrared spectra of Cortex Periplocae and Cortex Acanthopanacis was used for computation of the 2D correlation spectra and comparison of the synchronous and asynchronous 2D correlation spectra.Results:Two of the analytes had similar spectral characteristics in one-dimensional near infrared while quite different in 2D correlation spectra.Cortex Periplocae and Cortex Acanthopanacis were identified visually using the 2D correlation spectra in region 5 600~4 700 cm-1and 7 200~6 600 cm-1.Conclusion:The results prove that 2D correlation spectra could enhance the resolution of near infrared spectra and increase the capacity of identification.The combination of 2D correlation spectroscopy and near infrared spectroscopy allow a new and convenient approach for evaluating the medicinal herbs.
文摘旨在建立基于二维相关红外光谱(2D IR spectra)的分析技术,以进行中药中是否掺杂化学药物的判别。以减肥药盐酸芬氟拉明(fenfluramine hydrochloride,FH)、盐酸西布曲明(sibutramine hydrochloride,SH)为待检化合物,确定二者同步相关谱中的特征相关峰。首先通过比较掺杂中药与盐酸芬氟拉明的同步谱中特征相关峰的相似性,初步对掺杂化合物盐酸芬氟拉明进行判定,而后利用掺杂中药异步谱能确认相关峰来源的特性,进一步确证盐酸芬氟拉明的存在;此外,根据掺杂中药同步谱中未出现与盐酸西布曲明同步谱相似的特征相关峰,可以直接判别其中未掺杂盐酸西布曲明。实验所建立的方法无需样品分离,速度快,成本低,为快速简便地进行复杂体系中是否存在化学药物的判别提供了一种新的可能的方法。