基于深度学习的大肠埃希菌和志贺菌拉曼光谱鉴别模型

Identification model of Raman spectra for Escherichia coli and Shigella species based on deep learning

目的探讨表面增强拉曼散射(surface-enhanced Raman scattering,SERS)联合深度学习在大肠埃希菌与志贺菌鉴别中的价值。方法收集从徐州医科大学附属医院临床分离的大肠埃希菌及志贺菌各10株,建立上述细菌的SERS光谱数据集,应用深度学习分支卷积神经网络(convolutional neural network,CNN)构建2种细菌的分类模型。结果 SERS联合深度学习可以准确区分高度相似的大肠埃希菌和志贺菌,鉴别精确度为100%。与传统细菌鉴定方法相比,鉴定周期大大缩短。结论 SERS联合深度学习在大肠埃希菌和志贺菌的快速鉴别领域具有极大的应用价值和潜力。

Objective To explore the use of surface-enhanced Raman scattering(SERS)combined with deep learning for identification of Escherichia coli and Shigella species.Methods A total of 10 Escherichia coli strains and 10 Shigella strains were collected from the Affiliated Hospital of Xuzhou Medical University to establish a Raman spectroscopy database.Through the convolutional neural network(CNN),the classification models of the two bacteria were constructed.Results In the current study,highly similar Escherichia coli and Shigella species were precisely distinguished through SERS combined with deep learning,with an identification accuracy of 100%.Compared with the traditional bacterial identification methods,the identification period was significantly shortened.Conclusions SERS combined with deep learning is greatly useful in rapid identification of Escherichia coli and Shigella species.