Department of Systems Biology, Technical University of Denmark, Building 223, DK-2800, Kgs Lyngby, Denmark.
Appl Microbiol Biotechnol. 2010 May;86(6):1745-59. doi: 10.1007/s00253-009-2412-y. Epub 2010 Feb 5.
Fed-batch cultivations of Streptomyces coelicolor, producing the antibiotic actinorhodin, were monitored online by multiwavelength fluorescence spectroscopy and off-gas analysis. Partial least squares (PLS), locally weighted regression, and multilinear PLS (N-PLS) models were built for prediction of biomass and substrate (casamino acids) concentrations, respectively. The effect of combination of fluorescence and gas analyzer data as well as of different variable selection methods was investigated. Improved prediction models were obtained by combination of data from the two sensors and by variable selection using a genetic algorithm, interval PLS, and the principal variables method, respectively. A stepwise variable elimination method was applied to the three-way fluorescence data, resulting in simpler and more accurate N-PLS models. The prediction models were validated using leave-one-batch-out cross-validation, and the best models had root mean square error of cross-validation values of 1.02 g l(-1) biomass and 0.8 g l(-1) total amino acids, respectively. The fluorescence data were also explored by parallel factor analysis. The analysis revealed four spectral profiles present in the fluorescence data, three of which were identified as pyridoxine, NAD(P)H, and flavin nucleotides, respectively.
分批培养深红红螺菌，产生抗生素放线紫红素，通过多波长荧光光谱法和尾气分析进行在线监测。建立了偏最小二乘法（PLS）、局部加权回归和多线性偏最小二乘法（N-PLS）模型，分别用于预测生物量和基质（水解酪蛋白）浓度。考察了荧光和气体分析仪数据的组合以及不同变量选择方法的效果。通过将两个传感器的数据组合，并分别使用遗传算法、区间偏最小二乘法和主变量法进行变量选择，获得了改进的预测模型。逐步变量消除法应用于三通道荧光数据，得到了更简单、更准确的 N-PLS 模型。使用留一法进行交叉验证来验证预测模型，最好的模型的交叉验证值的均方根误差分别为 1.02 g l(-1)生物量和 0.8 g l(-1)总氨基酸。还通过平行因子分析对荧光数据进行了探索。分析揭示了荧光数据中存在四种光谱特征，其中三种分别被鉴定为吡哆醇、NAD(P)H 和黄素核苷酸。
