基于光谱技术对雨生红球藻虾青素含量检测及其空间分布的研究

发布时间：2018-06-08 19:09

本文选题：雨生红球藻 + 虾青素　；参考：《浙江大学》2017年硕士论文

【摘要】：雨生红球藻内积累的虾青素具有很高商业价值。通过该藻生产的虾青素被广泛地添加至保健品,化妆品及着色剂等产品中,在医药工业、食品工业、农业养殖等多个领域行业具有重要应用意义。本论文研究了雨生红球藻虾青素的最佳积累条件,基于光谱技术研究了雨生红球藻虾青素定量检测及细胞内相关成分空间分布可视化。主要得到以下结论:(1)设计正交实验,得到本论文所用雨生红球藻虾青素的最佳积累组合,并探究雨生红球藻在不同胁迫条件下的内部色素变化规律。通过设计正交实验,将氮源、碳源及盐度作为因素进行正交分析,通过直观法与正交方差法,分析得出本实验所用雨生红球藻虾青素积累量的最佳条件组合,即硝酸钠:0.15g/L,醋酸钠:1.5g/L,氯化钠:0.1%w/v,该组合成为后续研究中雨生红球藻胁迫条件的参考。该部分内容还分析了雨生红球藻在胁迫过程中,藻体色素含量的变化情况。随着胁迫时间的持续,叶绿素含量降低,虾青素不断的积累,类胡萝卜素总量也随之在不断增加。(2)应用可见/近红外光谱技术及高光谱成像技术获取雨生红球藻藻液光谱信息,将光谱信息与虾青素含量建立相应的定量模型,实现雨生红球藻虾青素含量快速检测;并发现高光谱成像技术建立的模型要优于可见/近红外光谱技术。对于可见/近红外光谱系统,通过分析获取198个样本在425～750nm波段范围吸收光谱与虾青素含量建立的定量模型,发现竞争性自适应加权采样算法(CARS)结合多元线性回归(MLR)建立的模型(CARS-MLR)最优,建模集相关系数Rcal达到0.976,验证集相关系数Rval达到0.961,预测集相关系数Rpre为0.967,建模均方根误差RMSEC为0.473,验证均方根误差RMSECV为0.606,预测均方根误差RMSEP为0.510,预测残余偏差RPD达到了 3.796。CARS-MLR模型也要优于使用全波段变量建立的偏最小二乘(PLS)模型,且前者的变量数仅为后者的2.39%。因此,425～750nm波段范围的吸收光谱能够实现雨生红球藻藻液中的虾青素含量检测,且通过CARS选取特征波段建立的CARS-MLR模型为最优,并得到相对应的定量模型公式。对于高光谱成像系统,通过分析获取198个样本在425～1023nm波段范围的反射光谱与虾青素含量建立的定量模型,发现竞争性自适应加权采样算法(CARS)结合偏最小二乘算法(PLS)建立的模型(CARS-PLS)最优,Rcal 达到 0.978,Rval 达到 0.967,Rpre 为 0.983,RMSEC为 0.443,RMSECV 为 0.544,RMSEP 为 0.398,RPD 达到了 5.324。CARS-PLS模型也要优于使用全波段变量建立的PLS模型,且前者的变量数仅为后者的6.75%。因此,425～1023nm波段范围的高光谱可见/近红外反射光谱能够实现雨生红球藻藻液虾青素含量检测,且通过CARS特征变量选取方法建立的CARS-PLS模型为最优,并得到雨生红球藻虾青素定量模型公式。通过比较两种技术建模结果,发现高光谱成像技术优于可见/近红外光谱技术。(3)利用共聚焦显微拉曼光谱技术对雨生红球藻类胡萝卜素及虾青素成分空间分布可视化进行了可行性研究。对比雨生红球藻拉曼光谱与虾青素标准品拉曼光谱的区别,分析出利用C=C拉伸键位移处的拉曼强度作为类胡萝卜素主要指标,得到雨生红球藻类胡萝卜素分布图;利用多元曲线分辨及交替最小二乘算法(MCR-ALS)分辨解析出雨生红球藻拉曼光谱中混杂着的虾青素信息,结合这些信息实现雨生红球藻虾青素在细胞内的空间分布的可视化。
[Abstract]:Astaxanthin accumulated in the Rhodococcus sp. is of high commercial value. The astaxanthin produced by this alga is widely added to health products, cosmetics and colorants and other products in the pharmaceutical industry, the food industry, the agricultural culture and other fields. The best product of the astaxanthin is studied in this paper. The quantitative detection of astaxanthin and the spatial distribution of related components in the cells were studied based on spectral techniques. The following conclusions were obtained: (1) the optimum accumulation of astaxanthin in this paper was obtained, and the internal pigmentation under different stress conditions was investigated. Through orthogonal analysis of nitrogen source, carbon source and salinity, the optimum conditions for the accumulation of astaxanthin in this experiment were obtained by orthogonal experiments. The optimum conditions for the accumulation of astaxanthin in this experiment were obtained, namely, sodium nitrate: 0.15g/L, sodium acetate: 1.5g/L, sodium chloride: 0.1%w/v, which became a follow-up study. This part also analyzed the changes of the content of the algal pigment during the stress process. With the persistence of the stress time, the content of chlorophyll decreased, the astaxanthin accumulated and the total carotenoid total increased. (2) the application of visible / near infrared spectroscopy and hyperspectral formation was applied. To obtain the spectral information of the Rhodococcus alga solution, establish the quantitative model of the spectral information and astaxanthin content, realize the rapid detection of the astaxanthin content of the rhodozina, and find that the hyperspectral imaging technology is better than the visible / near infrared spectroscopy. For the visible / near infrared spectral system, the 198 is obtained by analysis. A quantitative model of the absorption spectrum and astaxanthin content in the range of 425 ~ 750nm band was found, and the model (CARS-MLR) established by competitive adaptive weighted sampling algorithm (CARS) combined with multiple linear regression (MLR) was found to be the best. The correlation coefficient of the modeling set was 0.976, the correlation coefficient of the test set was 0.961, the correlation coefficient of the prediction set was 0, and the correlation coefficient Rpre of the prediction set was 0. .967, the root mean square error RMSEC is 0.473, the root mean square error RMSECV is 0.606, the predicted root mean square error RMSEP is 0.510, the predicted residual deviation RPD reaches the 3.796.CARS-MLR model and is superior to the partial least squares (PLS) model established by the full band variable, and the former is only the 2.39%. of the latter, and the 425 to 750nm band. The absorption spectrum of the range can be used to detect the astaxanthin content in the Rhodococcus alga solution, and the CARS-MLR model established in the characteristic band of CARS is the best, and the corresponding quantitative model formula is obtained. For hyperspectral imaging system, the reflectance spectra of 198 samples in the range of 425 ~ 1023nm band are obtained by analysis and astaxanthin. The quantitative model established by the content found that the competitive adaptive weighted sampling algorithm (CARS) combined with the partial least squares (PLS) model (CARS-PLS) is optimal, Rcal reaches 0.978, Rval reaches 0.967, Rpre is 0.983, RMSEC is 0.443, RMSECV is 0.544, RMSEP is 0.398, 5.324.CARS-PLS model is also better than the use of full wave. The PLS model is established by the segment variable, and the former is only 6.75%. of the latter, so the hyperspectral visible / near infrared reflectance spectrum of the 425 ~ 1023nm band can be used to detect the astaxanthin content in the alga alga solution. And the CARS-PLS model established by the CARS characteristic variable selection method is the best, and the astaxanthin in the raining red red alga is obtained. Quantitative model formula. By comparing the results of two technical modeling, it is found that hyperspectral imaging technology is superior to visible / near infrared spectroscopy. (3) the feasibility of visualization of the spatial distribution of carotene and astaxanthin components in raining red ball, alga, carotene and astaxanthin is studied by using confocal Raman microraman spectroscopy. The Raman spectra of the standard products were distinguished, and the Raman intensity of the C=C tensile bond displacement was analyzed as the main index of carotenoid. The distribution diagram of the carotene of the raining red ball algae was obtained. The information of the astaxanthin mixed in the Raman spectrum of the Rhodococcus Rhodococcus was resolved by using the multiple curve resolution and the alternate least squares algorithm (MCR-ALS). Combined with these information, we can visualize the spatial distribution of astaxanthin in the cell.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：Q946

【相似文献】