摘要
目的:利用反向传播(Back propagation,BP)神经网络结合遗传算法(Genetic Algorithm,GA)建立预测模型,拟合链霉菌发酵条件与酶产量的关系,为生产葡萄糖异构酶提供最优发酵条件。方法:通过单因素试验优化葡萄糖异构酶的发酵条件,采用正交试验考察酵母膏添加量、初始pH值、接种量对酶产量的影响。利用GA在试验水平范围预测全局最优发酵条件。结果表明:BP神经网络的训练集、验证集、测试集和总体数据的相关系数分别为1、0.9999、1和1,表明BP神经网络预测模型的准确性很好,可用于链霉菌发酵葡萄糖异构酶酶产量的预测。结论:利用GA寻优的结果为50 mL液体培养基中加入0.3 g酵母膏,初始pH值为7,接种量5%的葡萄糖异构酶的酶产量为7.38 U/mL,与模型预测值的误差仅为1.76%,比初始发酵条件发酵液5.3 U/mL提高了41.1%,说明BP神经网络模型结合GA是一种可用于优化发酵条件提高葡萄糖异构酶产量的方法。
The method of the back propagation (BP) neural network combined with the genetic algorithm (GA) to establish the prediction model was used to find the relationship between the fermentation condition of Streptomyces and the yield of glucose isomerase,which provided the optimal fermentation condition for producing glucose isomerase.The fermentation condition of glucose isomerase was optimized through the single factor experiment.Orthogonal experiment was and carried out to investigate the effect of the added amount of yeast extract,initial pH and inoculums on the yield of glucose isomerase.The GA was used to predict the overall optimal fermentation condition in the experimental level.The result showed that the correlation coefficients of BP neural network training,verification,testing and all the data were 1,0.9999,1 and 1,respectively,indicating that BP neural network prediction model had good accuracy and could be used to predict the yield of glucose isomerase which was got through the fermentation of Streptomyces.The results of using GA for optimization were yeast extract 0.3 g,initial pH 7,inoculums 5%.The experimental value of the yield of glucose Isomerase was 7.38 U/mL,and the error of the model predicted value was only 1.76%.This was 41.1% higher than the yield of original enzyme of 5.3 U/mL.The results indicated BP neural network combined with GA was a method with better accuracy to optimize the fermentation condition to raise the yield of glucose isomerase.
作者
魏泉增
范江涛
张成丽
孙军涛
张永清
黄继红
WEI Quanzeng;FAN Jiangtao;ZHANG Chengli;SUN Juntao;ZHANG Yongqing;HUANG Jihong(Key Laboratory of Biomarker Based Rapid-Detection Technology for Food Safety of Henan Province,Food and Pharmacy College,Xuchang University,Xuchang 461000,China)
出处
《食品科技》
CAS
北大核心
2021年第12期34-41,共8页
Food Science and Technology
基金
中原院士基金·中原学者项目(192101510004)
国家绿色制造2025项目(工信部节函[2018]272号-3)
中国工程科技发展战略河南研究院战略咨询研究项目(2020HENZT13)。
关键词
葡萄糖异构酶
正交试验
神经网络
遗传算法
glucose isomerase
orthogonal experiment
neural network
genetic algorithm
