基于神经网络和遗传算法的岩藻多糖酶发酵培养基优化  

Optimization of Fermentation Medium of Fucoidanase Based on Artificial Neural Network and Genetic Algorithm

作　　者：王莹[1] 仇宏伟[1] 

机构地区：[1]青岛农业大学食品科学与工程学院,山东青岛266109

出　　处：《青岛农业大学学报（自然科学版）》2014年第2期131-135,共5页Journal of Qingdao Agricultural University(Natural Science)

基　　金：青岛农业大学高层次人才启动基金(631347)

摘　　要：本文对RC2-3产岩藻多糖酶的发酵培养基进行优化。通过单因素实验确定了最佳碳源、氮源、无机盐种类,并利用正交实验及人工神经网络与遗传算法联合优化确定了培养基浓度。结果表明最佳碳、氮源分别为岩藻多糖、牛肉膏,最佳无机盐为NaCl。利用正交实验优化得到了最佳培养基浓度为岩藻多糖0.6%,牛肉膏0.8%,NaCl5%,在该条件下可以获得酶活力68.8±3.2U/mL;进一步利用神经网络与遗传算法联合优化培养基浓度,得到最优组合为岩藻多糖0.54%,牛肉膏0.40%,NaCl6.66%,此条件下能够获得酶活力78.4±5.3U/mL。利用神经网络与遗传算法联合优化的结果比优化前酶活力提高了45%,比正交实验优化结果提高了14%,证明该方法在培养基优化中具有显著优越性。In this paper, the optimized medium of fucoidanase produced by bacterium RC2-3 was obtained by single factor experiment to determine the best carbon source, nitrogen source, and inorganic salt, then using orthogonal experiment and artificial neural network combined genetic algorithm to determine the best medium concenration. The results showed that the best carbon source is fucoidan, the best nitrogen source is beef extract, and the best inorganic salt is NaCl. The optimized medium was obtained by a orthogonal experiment, which contained fucoidan 0.6 Y00, beef extract 0.8 %, NaCl 5 %, and in this condition enzyme activity is 68.8±3.2U/mL. Furthermore, using the artificial neural network combined the genetic algorithm, the optimal combination is fucoidan 0.54%, beef extract 0.40%, NaCl 6.66%, in this condition enzyme activity is 78.4±5.3U/mL. The result indicated that the enzyme activity of the optimized fermen- tation medium obtained by GA and ANN was increased by 45% than the original medium and 14% than the medium obtained by the orthogonal experiment. The combination of artificial neural network and genetic algorithm showed a great advantage in medium optimization.

关 键 词：岩藻多糖酶 发酵 培养基 神经网络 遗传算法 

分 类 号：Q814[生物学—生物工程]