液相色谱和多变量分析鉴别结核分枝杆菌菌株

Discriminating Mycobacterial Strains by LC and Multivariate Analysis

Abstract

Traditional methods of mycobacterial identification relied on a battery of biochemical tests. Most of these tests were low resolution—yes or no, for example—and could be subjective. Chromatographic procedures have replaced these tests in many laboratories because of ease of implementation and greater objectivity in the interpretation of the results.

Liquid chromatography (LC) of bacterial extracts has been demonstrated as a useful technique to separate the high molecular weight mycolic acids that are species specific. Multivariate analysis of the LC profiles can help the analyst distinguish among the different bacterial species. Tuberculosis is caused by one species of Mycobacteria—there are other species in this genus which are also human pathogens—thus reliable identification of species is critical.

Conclusions

This application note provides both a strategy and a procedure for extracting the information content from large, complex
chromatography databases.  If the data is aligned to remove retention time variation, principal component analysis is successful in organizing the data into natural groupings. Using an interactive graphical display of PCA scores provides a means to identify samples of interest, where the overlap density map can take over to highlight portions of the chromatogram that reflect similarities and differences.