Discrimination of Microbe Species by Laser Induced Breakdown Spectroscopy
10.11895/j.issn.0253-3820.171448
- VernacularTitle:基于激光诱导击穿光谱的微生物种类鉴别研究
- Author:
Gang-Fu RAO
1
;
Lin HUANG
;
Mu-Hua LIU
;
Tian-Bing CHEN
;
Jin-Yin CHEN
;
Zi-Yi LUO
;
Fang-Hao XU
;
Hui YANG
;
Xiu-Wen HE
;
Hua-Mao ZHOU
;
Jin-Long LIN
;
Ming-Yin YAO
Author Information
1. 江西农业大学工学院
- Keywords:
Microbes;
Rapid identification;
Plasma plume;
Laser induced breakdown spectroscopy;
Random forest;
Principal component analysis
- From:
Chinese Journal of Analytical Chemistry
2018;46(7):1122-1128
- CountryChina
- Language:Chinese
-
Abstract:
Laser induced breakdown spectroscopy ( LIBS ) was proposed to rapidly discriminate microbe species. Ten species of microbes were prepared in lab. Filter papers were selected as substrate for enriching bacteria and enhancing the quality of LIBS. The images of plasma were collected by ICCD camera and LIBS spectra were obtained by spectrometers. The results displayed that the images and spectra were different from 10 bacteria. It was demonstrated that this method was feasible to discriminate bacteria species by analyzing image and/or spectroscopy. Furthermore, nine smooth and multiple scattering correction ( MSC) were utilized to preprocess the LIBS full-spectrum data in the wavelength range of 200-420 nm and 560-680 nm. And principal component analysis ( PCA) and PCA-RF ( Random forest) were compared to validate the accuracy of discrimination. The investigation showed that the PCA-RF model coupled with suitable methods in preprocessing data could identify bacteria. The accuracy was 99. 6% for ten species of microbes by evaluating LIBS spectra in training set, and 96. 7% in predicting set. This report indicated that it is feasible to differentiate bacteria species by analyzing LIBS spectra.
