Variable selection and multivariate methods for the identification of microorganisms by flow cytometry

Background: When exploited fully, flow cytometry can be used to provide mul tiparametric data for each cell in the sample of interest. While this makes flow cytometry a powerful technique for discriminating between different c ell types, the data can be difficult to interpret. Traditionally, dual-para meter plots are used to visualize now cytometric data, and for a data set c onsisting of seven parameters, one should examine 21 of these plots. A more efficient method is to reduce the dimensionality of the data (e.g., using unsupervised methods such as principal components analysis) so that fewer g raphs need to be examined, or to use supervised multivariate data analysis methods to give a prediction of the identity of the analyzed particles. Materials and Methods: We collected multiparametric data sets for microbiol ogical samples stained with six cocktails of fluorescent stains. Multivaria te data analysis methods were explored as a means of microbial detection an d identification. Results: We show that while all cocktails and all methods gave good accurac y of predictions (>94%), careful selection of both the stains and the analy sis method could improve this figure (to >99% accuracy), even in a data set that was not used in the formation of the supervised multivariate calibrat ion model. Conclusions: Flow cytometry provides a rapid method of obtaining multiparam etric data for distinguishing between microorganisms. Multivariate data ana lysis methods have an important role to play in extracting the information from the data obtained. Artificial neural networks proved to be the most su itable method of data analysis. (C) 1999 Wiley Liss, Inc.