目前随着人工智能的快速发展,机器学习已经广泛应用于各个领域,并对实验中获得的小数据进行模拟分析。本研究以微藻为基础,利用收集的9组文献中对微藻的研究结果为数据,使用四种算法,即BP神经网络、支持向量机、随机森林和径向基函数神经网络,使用MATLAB软件进行建模和分析,与原文献中的响应面分析法进行对比,得出以下主要结论:通过比较相对系数(R2)、平均绝对误差(MAE)、平均偏差误差(MBE)、均方根误差(RMSE)和均方误差(MSE),相比于传统的响应面分析法,机器学习算法表现出更好的预测效果,其中随机森林和径向基函数神经网络的相对系数最接近于1,预测效果最好,其次是BP神经网络和支持向量机。Currently, with the rapid development of artificial intelligence, machine learning has been widely applied in various fields and used to simulate and analyze small data obtained from experiments. This study based on microalgae and using the collected 9 groups of microalgae research results in the literature for the data, using four kinds of algorithms, namely, the BP neural network, support vector machine (SVM), random forests (RF) and radial basis function (RBF) neural network, using MATLAB software for modeling and analysis, comparing with the original documents by response surface analysis method. The following main conclusions: by comparing the relative coefficient (R2), the mean absolute error (MAE), mean bias error (MBE), root mean square error (RMSE) and mean square error (MSE), compared with the traditional response surface analysis method, machine learning algorithms show better prediction effect, random forests and the relative coefficient of radial basis function (RBF) neural network is the most close to 1, the prediction effect is best, followed by BP neural network and support vector machine.
本研究从球等鞭金藻(Isochrysis galbana)藻际环境中分离获得一株耐盐碱菌株Microbacterium paraoxydans IGS-10,通过生理生化特性分析、16S rRNA系统发育学鉴定及全基因组测序技术,系统解析其耐盐碱分子机制。实验结果显示,IGS-10在NaCl浓度≤6% (w/v)和pH 7-10范围内均能稳定生长,表明该菌株对高盐碱环境具有显著适应性。基因组测序揭示其基因组大小约为3.84 Mb,GC含量70.35%,并携带30余个与耐盐碱密切相关的功能基因,包括离子转运系统基因以及相容性溶质(甜菜碱和海藻糖)合成基因。本研究不仅丰富了藻际微生物资源库,还为解析极端环境微生物的适应性进化机制提供了关键分子证据,对耐盐碱微生物资源的农业与工业应用具有潜在参考价值。In this study, a saline-tolerant strain Microbacterium paraoxydans IGS-10 has been isolated from Isochrysis galbana in the algal environment. Through physiological and biochemical analysis, 16S rRNA phylogenetic identification and whole genome sequencing technology, Microbacterium Paraoxydans IGS-10 has been obtained. The molecular mechanism of saline-alkali resistance was systematically analyzed. The experimental results showed that IGS-10 could grow stably in the range of NaCl concentration ≤6% (w/v) and pH 7-10, indicating that the strain had remarkable adaptability to high saline-alkali environment. Genome sequencing revealed that its genome size was about 3.84 Mb, GC content was 70.35%, and it carried more than 30 functional genes closely related to saline-alkali tolerance, including ion transport system genes and compatible soles (betaine and trehalose) synthesis genes. This study not only enriches the algal microbial resource pool, but also provides key molecular evidence for the analysis of the adaptive evolutionary mechanism of microorganisms in extreme environments, and has potential reference value for the agricultural and industrial application of saline-tolerant microbial resources.
