Multivariate unmixing approaches on Raman images of plant cell walls: new insights or overinterpretation of results?

Abstract

BackgroundPlant cell walls are nanocomposites based on cellulose microfibrils embedded in a matrix of polysaccharides and aromatic polymers. They are optimized for different functions (e.g. mechanical stability) by changing cell form, cell wall thickness and composition. To reveal the composition of plant tissues in a non-destructive way on the microscale, Raman imaging has become an important tool. Thousands of Raman spectra are acquired, each one being a spatially resolved molecular fingerprint of the plant cell wall. Nevertheless, due to the multicomponent nature of plant cell walls, many bands are overlapping and classical band integration approaches often not suitable for imaging. Multivariate data analysing approaches have a high potential as the whole wavenumber region of all thousands of spectra is analysed at once.ResultsThree multivariate unmixing algorithms, vertex component analysis, non-negative matrix factorization and multivariate curve resolution–alternating least squares were applied to find the purest components within datasets acquired from micro-sections of spruce wood and Arabidopsis. With all three approaches different cell wall layers (including tiny S1 and S3 with 0.09–0.14 µm thickness) and cell contents were distinguished and endmember spectra with a good signal to noise ratio extracted. Baseline correction influences the results obtained in all methods as well as the way in which algorithm extracts components, i.e. prioritizing the extraction of positive endmembers by sequential orthogonal projections in VCA or performing a simultaneous extraction of non-negative components aiming at explaining the maximum variance in NMF and MCR-ALS. Other constraints applied (e.g. closure in VCA) or a previous principal component analysis filtering step in MCR-ALS also contribute to the differences obtained. ConclusionsVCA is recommended as a good preliminary approach, since it is fast, does not require setting many input parameters and the endmember spectra result in good approximations of the raw data. Yet the endmember spectra are more correlated and mixed than those retrieved by NMF and MCR-ALS methods. The latter two give the best model statistics (with lower lack of fit in the models), but care has to be taken about overestimating the rank as it can lead to artificial shapes due to peak splitting or inverted bands. Electronic supplementary materialThe online version of this article (10.1186/s13007-018-0320-9) contains supplementary material, which is available to authorized users.

Analysis of Cellulose and Lignocellulose Materials by Raman Spectroscopy: A Review of the Current Status

“…For example, PLS was used to build a predictive model for estimating cellulose crystallinity for cellulose materials [23]. Other methods were subsequently applied to plant cell walls [15,54,55,56]; complex multicomponent lignocellulose samples, that produced complex Raman spectra so that the identification and quantitation of individual constituents can be carried out. To automatically identify Raman spectra of different cell wall layers (cell corner—CC, compound middle lamella—CML, secondary wall—SW, gelatinous layer—G-layer, and cell lumen), Zhang et al [54] proposed a new chemometric method on the basis of PCA and cluster analysis.…”

From the Soft to the Hard: Changes in Microchemistry During Cell Wall Maturation of Walnut Shells

“…Zooming into the inner part of the July sample ( Figure 6A) and applying a multivariate unmixing approach, the onset of lignification was visualized in detail ( Figures 6B-E). Non negative matrix factorization (NMF) delivers the most pure component (endmember) spectra and their abundance maps (Prats-Mateu et al, 2018). One component was retrieved from CC, CML and pits ( Figure 6B) and showed clear aromatic bands at 1,657, 1,598, 1,335, 1,140 cm −1 together with the pectin band at 853 cm −1 (Synytsya et al, 2003) (Figure 6D, magenta).…”

Raman imaging of protein in a model cheese system

“…Unsupervised MCR-ALS is commonly used to analyse spectral image data; in this case, it proved to differentiate the three components of interest in the cheese matrix easily by applying a minimum of three end members to the algorithm. 16,17 Although smaller fat globules appear to be contained within the casein component map instead of the fat component map, as in Figure 8, this could possibly be resolved by further chemometric techniques that were not investigated in this study. Examination of the spectra show the complete absence of water signals in the region 3100-3550 cm -1 in the fat (red) spectra along with indicative carbonyl signal from the fat triglyceride at 1740 cm -1 .…”