基于主成分分析的食源性致病菌的投影判别分析

Projection Discriminant Analysis of Food-borne Pathogens Based on Principal Component Analysis

目的:以大肠杆菌、单增李斯特菌和沙门氏菌为研究对象,建立近红外光谱检测食源性致病菌的方法。方法:利用近红外光谱分析技术,对大肠杆菌、单增李斯特菌和沙门氏菌的细胞壁、细胞质和全细胞分别进行基于主成分分析的投影判别分析。结果:多元散射校正法对光谱预处理比矢量归一法预处理后的细胞质壁分离更为有效;采用细胞壁进行判别比用细胞质判别3种致病菌的结果好;G阴性菌与G阳性菌之间的判别比同性菌之间判别准确度高,对细胞壁的分辨率达到100%。结论:利用基于近红外光谱主成分分析的投影判别分析技术,可以检测和鉴别上述3种致病菌,且其分辨率为100%。

Object：Take Escherichia coli（E.coli）, Listeria monocytogenes（LM） and salmonella as examples, the Near Infrared Spectroscopy method of identifying food-borne pathogens was established. Methods：Using Near Infrared Spectroscopy technique, the cell wall, the cytoplasm and the whole cell of E.coli, LM and salmonella were analyzed separately by projection discriminant analysis based on principal component analysis. Results： The separating effect of cell wall with cytoplasm pretreated by multiple scattering correction was better than that by vector normalization; The result of cell wall was optimal; The accuracy of detection between G＋ and G- was higher than the same gram properties, and the resolution of cell wall reached to 100%. Conclusions：The three kinds of food-borne pathogen could be detected and identified by the projection discriminant analysis based on principal component analysis and near infrared spectroscopy technique, and their resolutions were 100%.