Application of cluster analysis (CLA) in feed chemical imaging to accurately reveal structural-chemical features of feeds and plants within cellular dimension

Synchrotron Fourier transform infrared (FTIR) microspectroscopy can explore molecular chemical features of the microstructure of feeds. The most straightforward method of data analysis is the mapping of specific functional group intensities and frequencies by peak heights and/or areas. However, this univariate statistical method does not always accurately identify functional group locations and concentrations, because the so-called "unique" bands for the peak area mapping have more or less inference with other nonunique bands in feed and plant tissues. The objective of this study was to use a multivariate analysis method--called agglomerative hierarchical cluster analysis (CLA)--to analyze infrared spectra for chemical imaging. The results show the CLA method gave satisfactory results and was conclusive in showing that it can discriminate and classify functional group differences existing in different structure regions. This approach (CLA for chemical imaging) places synchrotron FTIR microspectroscopy at the forefront of those techniques that could potentially be used in the rapid characterization of feed microstructure.