Posthandling Spectral Information Enhancement for Single Cell Raman Molecular Mapping Analysis.

Biochemical analysis of living systems such as single cells benefits greatly from the label-free and low-invasive molecular mapping with Raman microspectroscopy. Sets of Raman spectra at different spatial points are analyzed to generate Raman molecular maps corresponding to specific chemical species. However, human error and subjective data analysis can be technical issues that limit interpretation and its validity. Here, we present an objective data analysis scheme for postprocessing large data sets of Raman spectra for molecular mapping of living cells. The process comprises three steps: (i) Denoising the spectral data set using low-rank approximation; (ii) obtaining an objective background from data points outside the target cell; (iii) subtracting the thus obtained background using Hypothetical Addition Multivariate Analysis with Numerical Differentiation (HAMAND) via an automatically determined coefficient. Through the present analysis, minor Raman peaks, as indiscernible as they are, can be identified and precisely mapped. We demonstrate a quantitative discussion of cellular components after extracting contributions only from a single target cell from a Raman mapping image where multiple cells or parts of other cells are present. This work opens an improved analysis workflow for accurate spectroscopic analysis of living cells with the advantage of identifying minor Raman peaks unambiguously.