谷氨酸棒杆菌发酵过程中亮氨酸浓度近红外模型的建立  

Establishment of Near-Infrared Model of Leucine Concentration in Corynebacterium glutamicum Fermentation Process

作　　者：马雷[1,2] 郑梦鸽 石拓[3] 马跃超[3] 谢希贤[3] 陈宁[3] 

机构地区：[1]天津职业技术师范大学,天津300222 [2]天津科技大学电子信息与自动化学院,天津300457 [3]天津科技大学生物工程学院,天津300457

出　　处：《生物技术通讯》2016年第4期501-506,共6页Letters in Biotechnology

基　　金：天津市应用基础与前沿技术研究计划(14JCYBJC23500)

摘　　要：目的：建立谷氨酸棒杆菌发酵过程中亮氨酸浓度近红外模型,为实现谷氨酸棒杆菌发酵生产亮氨酸的发酵过程自动化控制提供理论基础和实践依据。方法：首先在5 L发酵罐中进行亮氨酸发酵,每隔一段时间采集发酵液样品,用高效液相色谱精确分析各样品中的亮氨酸实际浓度,再利用近红外分析仪和相关软件,对各样品进行近红外光谱扫描分析,并通过近红外光谱分析软件进行数据处理,建立谷氨酸棒杆菌发酵过程中亮氨酸浓度的近红外预测模型,最后通过外部检验方法检验模型的准确性。结果：结合偏最小二乘法,在波长为9043.3~7489.1 cm-1、减去一条直线作为光谱预处理的条件下,获得谷氨酸棒杆菌发酵过程中亮氨酸浓度最优近红外预测模型。该模型交叉验证误差均方根（RMSECV）、决定系数（R2）以及剩余预测偏差（RPD）分别为1.29 g/L、0.977和4.55。结论：经过验证,该模型的准确性和可靠性较强,实际值与预测值之间的误差较小,能够较好地检测发酵过程中的亮氨酸浓度。Objective： To establish a near-infrared model of leucine concentration in the fermentation process of Corynebacterium glutamicun. And to provide theoretical and practical basis for the automatic control of fermenta-tion process. Methods： Firstly, the leucine fermentation was performed in the 5 L fermentation tank. Cultureswere collected during fermentation period according to a certain interval. Secondly, the leucine concentration of cul-ture samples was accurately detected with high performance liquid chromatography. Then, the near-infrared spec-trum scanning analysis was carried out by using near infrared analyzer and related software, and the predictionmodel of leucine concentration was established with the near-infrared analysis software. Finally, the accuracy andreliability of the model were tested by the external inspection method. Results： The near- infrared spectroscopywas used combining with partial least squares method to get the best prediction model under the condition of mi-nus a straight line for pretreatment and 9043.3~7489.1 cm- 1wavelength. The value of root mean square error ofcross validation（RMSECV）, determination coefficient（R2）, and residual prediction error（RPD） was 1.29 g/L, 0.977,and 4.55, respectively. Conclusion： In the confirmatory experiment, the model showed high accuracy, where the er-ror between the actual value and the predicted value was low, and could predicted the concentration of leucine re-liability in the fermentation process.

关 键 词：近红外 预测模型 亮氨酸 谷氨酸棒杆菌 

分 类 号：TQ922[轻工技术与工程—发酵工程]