Search

Principal Component Analysis (PCA) of primary human T-cell spectra showed clear separation of different fixatives. Saline dry-fixed cell spectra showed better preservation of donor-to donor variability and improved retention of Raman band shapes.

Two-dimensional correlation spectroscopy (2D-COS) is a technique that permits the examination of synchronous and asynchronous changes present in hyperspectral data. It produces two-dimensional correlation coefficient maps that represent the mutually correlated changes occurring at all Raman wavenumbers during an implemented perturbation. To focus our analysis on clusters of wavenumbers that tend to change together, we apply a k-means clustering to the wavenumber profiles in the perturbation domain decomposition of the two-dimensional correlation coefficient map. These profiles (or trends) reflect peak intensity changes as a function of the perturbation. We then plot the co-occurrences of cluster members two-dimensionally in a manner analogous to a two-dimensional correlation coefficient map. Because wavenumber profiles are clustered based on their similarity, two-dimensional cluster member spectra reveal which Raman peaks change in a similar manner, rather than how much they are correlated. Furthermore, clustering produces a discrete partitioning of the wavenumbers, thus a two-dimensional cluster member spectrum exhibits a discrete presentation of related Raman peaks as opposed to the more continuous representations in a two-dimensional correlation coefficient map. We demonstrate first the basic principles of the technique with the aid of synthetic data. We then apply it to Raman spectra obtained from a polystyrene perchlorate model system followed by Raman spectra from mammalian cells fixed with different percentages of methanol. Both data sets were designed to produce differential changes in sample components. In both cases, all the peaks pertaining to a given component should then change in a similar manner. We observed that component-based profile clustering did occur for polystyrene and perchlorate in the model system and lipids, nucleic acids, and proteins in the mammalian cell example. This confirmed that the method can translate to “real world” samples. We contrast these results with two-dimensional correlation spectroscopy results. To supplement interpretation, we present the cluster-segmented mean spectrum of the hyperspectral data. Overall, this technique is expected to be a valuable adjunct to two-dimensional correlation spectroscopy to further facilitate hyperspectral data interpretation and analysis.

Here, we present an augmented form of two-dimensional correlation spectroscopy, that integrates in a single format data from spectroscopic and multiple non-spectroscopic sources for analysis. The integration is affected by augmenting every spectrum in a hyperspectral data set with relevant non-spectroscopic data to permit two-dimensional correlation analysis(2D-COS) of the ensemble of augmented spectra. A k-means clustering is then applied to the results of the perturbation domain decomposition to determine which Raman peaks cluster with any of the non-spectroscopic data. We introduce and explain the method with the aid of synthetic spectra and synthetic non-spectroscopic data. We then demonstrate this approach with data using Raman spectra from human embryonic stem cell aggregates undergoing directed differentiation toward pancreatic endocrine cells and parallel bioassays of hormone mRNA expression and C-peptide levels in spent medium. These pancreatic endocrine cells generally contain insulin or glucagon. Insulin has disulfide bonds that produce Raman scattering near 513 cm–1, but no tryptophan. For insulin-positive cells, we found that the application of multisource correlation analysis revealed a high correlation between insulin mRNA and Raman scattering in the disulfide region. In contrast, glucagon has no disulfide bonds but does contain tryptophan. For glucagon-positive cells, we also observed a high correlation between glucagon mRNA and tryptophan Raman scattering (∼757 cm–1). We conclude with a discussion of methods to enhance spectral resolution and its effects on the performance of multisource correlation analysis.

Spectroscopic peak parameters are important since they provide information about the analyte under study. Besides obtaining these parameters, peak fitting also resolves overlapped peaks. Thus, the obtained parameters should permit the construction of a higher-resolution version of the original spectrum. However, peak fitting is not an easy task due to computational reasons and because the true nature of the analyte is often unknown. These difficulties are major impediments when large hyperspectral data sets need to be processed rapidly, such as for manufacturing process control. We have developed a novel and relatively fast two-part algorithm to perform peak fitting and resolution enhancement on such data sets. In the first part of the algorithm, estimates of the total number of bands and their parameters were obtained from a representative spectrum in the data set, using a combination of techniques. Starting with these parameter estimates, all the spectra were then iteratively and rapidly fitted with Gaussian bands, exploiting intrinsic features of the Gaussian distribution with vector operations. The best fits for each spectrum were retained. By reducing the obtained bandwidths and commensurately increasing their amplitudes, high-resolution spectra were constructed that greatly improved correlation-based analyses. We tested the performance of the algorithm on synthetic spectra to confirm that this method could recover the ground truth correlations between highly overlapped peaks. To assess effective peak resolution, the method was applied to low-resolution spectra of glucose and compared to results from high-resolution spectra. We then processed a larger spectral data set from mammalian cells, fixed with methanol or air drying, to demonstrate the resolution enhancement of the algorithm on complex spectra and the effects of resolution-enhanced spectra on two-dimensional correlation spectroscopy and principal component analyses. The results indicated that the algorithm would allow users to obtain high-resolution spectra relatively fast and permit the recovery of important aspects of the data's intrinsic correlation structure.

Overlapping peaks in Raman spectra complicate the presentation, interpretation, and analyses of complex samples. This is particularly problematic for methods dependent on sparsity such as multivariate curve resolution and other spectral demixing as well as for two-dimensional correlation spectroscopy (2D-COS), multisource correlation analysis, and principal component analysis. Though software-based resolution enhancement methods can be used to counter such problems, their performances often differ, thereby rendering some more suitable than others for specific tasks. Furthermore, there is a need for automated methods to apply to large numbers of varied hyperspectral data sets containing multiple overlapping peaks, and thus methods ideally suitable for diverse tasks. To investigate these issues, we implemented three novel resolution enhancement methods based on pseudospectra, over-deconvolution, and peak fitting to evaluate them along with three extant methods: node narrowing, blind deconvolution, and the general-purpose peak fitting program Fityk. We first applied the methods to varied synthetic spectra, each consisting of nine overlapping Voigt profile peaks. Improved spectral resolution was evaluated based on several criteria including the separation of overlapping peaks and the preservation of true peak intensities in resolution-enhanced spectra. We then investigated the efficacy of these methods to improve the resolution of measured Raman spectra. High resolution spectra of glucose acquired with a narrow spectrometer slit were compared to ones using a wide slit that degraded the spectral resolution. We also determined the effects of the different resolution enhancement methods on 2D-COS and on chemical contrast image generation from mammalian cell spectra. We conclude with a discussion of the particular benefits, drawbacks, and potential of these methods. Our efforts provided insight into the need for effective resolution enhancement approaches, the feasibility of these methods for automation, the nature of the problems currently limiting their use, and in particular those aspects that need improvement., Graphical Abstract

Therapies based on injecting living cells into patients offer a huge potential to cure many degenerative and deadly diseases, with hundreds of clinical trials ongoing. Due to their complex nature, a basic understanding of their biochemical and functional characteristics, how to manufacture them for safe and efficacious therapy, and how to effectively implement them in clinical settings are very challenging. Raman spectroscopy could provide an information-rich, non-invasive, non-destructive analytical method to complement the use of conventional sample-based, infrequent and destructive biochemical assays typically employed to analyze and validate the quality of therapeutic cells. This article provides an overview of the current state of emerging cell therapies, and then reviews the related Raman spectroscopic state of the art analysis of human cells. This includes spectroscopic data processing considerations, the scope offered by technical variants of Raman spectroscopy, and analytical difficulties encountered by spectroscopists working with therapeutic cells. Finally, we outline a number of salient challenges as cell therapy products are translated from the laboratory to the clinic, and propose how Raman spectroscopy-based solutions could address these challenges.

Supplemental material, sj-pdf-2-asp-10.1177_0003702820979331 for Augmented Two-Dimensional Correlation Spectroscopy for the Joint Analysis of Correlated Changes in Spectroscopic and Disparate Sources by H. Georg Schulze, Shreyas Rangan, Martha Z. Vardaki, Diepiriye G. Iworima, Timothy J. Kieffer, Michael W. Blades, Robin F. B. Turner and James M. Piret in Applied Spectroscopy

Supplemental material, sj-pdf-1-asp-10.1177_0003702820979331 for Augmented Two-Dimensional Correlation Spectroscopy for the Joint Analysis of Correlated Changes in Spectroscopic and Disparate Sources by H. Georg Schulze, Shreyas Rangan, Martha Z. Vardaki, Diepiriye G. Iworima, Timothy J. Kieffer, Michael W. Blades, Robin F. B. Turner and James M. Piret in Applied Spectroscopy

OBJECTIVES The approximately 250 years old remains of the Kwadąy Dan Ts'inchį man were found in a glacier in Canada. Studying the state of preservation of the corpse, we observed black deposits in his lung. Following this observation we wanted to determine: (1) location of the deposits in the lung tissue, (2) composition and origins of the deposits. METHODS By light microscopy (LM) and transmission electron microscopy (TEM), we studied the deposits in the Kwadąy Dan Ts'inchį man' s lung and compared it with distribution of anthracotic deposits in contemporary samples from the David Harwick Pathology Centre (DHPC). To determine chemical composition of the inclusions we used Raman spectroscopy. Scanning electron microscopy and elemental mapping was used for determine the chemical elements. RESULTS The histopathological identification of anthracosis in the Kwadąy Dan Ts'inchį man's lung allowed us to distinguish crushed parenchyma from conducting airway tissue and identification of particles using LM and TEM. Crystal particles were found using TEM. Ordered carbonaceous material (graphene and graphite), disordered carbonaceous material (soot) and what might be minerals (likely conglomerates) were found with Raman spectrometry. Gold and lead particles in the lung were discovered with scanning electron microscopy and elemental mapping. CONCLUSIONS Presence of soot particles in anthracotic areas in the Kwadąy Dan Ts'inchį man's lung probably were due to an inhalation of particles in open fires. Gold and lead particles are most likely of an environmental origin and may have been inhaled and could have impacted his health and his Champagne and Aishihik First Nations (CAFN) contemporaries.

Spectral preprocessing is frequently required to render Raman spectra useful for further processing and analyses. The various preprocessing steps, individually and sequentially, are increasingly being automated to cope with large volumes of data from, for example, hyperspectral imaging studies. Full automation of preprocessing is especially desirable when it produces consistent results and requires minimal user input. It is therefore essential to evaluate the “quality” of such preprocessed spectra. However, relatively few methods exist to evaluate preprocessing quality, and fully automated methods for doing so are virtually non-existent. Here we provide a brief overview of fully automated spectral preprocessing and fully automated quality assessment of preprocessed spectra. We follow this with the introduction of fully automated methods to establish figures-of-merit that encapsulate preprocessing quality. By way of illustration, these quantitative methods are applied to simulated and real Raman spectra. Quality factor and quality parameter figures-of-merit resulting from individual preprocessing step quality tests, as well as overall figures-of-merit, were found to be consistent with the quality of preprocessed spectra.

Blood banking is an essential aspect of modern medical care. When red blood cells (RBCs) are stored, they become damaged by various chemical processes, such as accumulation of their own waste products and oxidative injury, among others. These processes lead to the development of the RBC storage lesion, a complex condition where the severity is reflected through the morphology of the stored cells. It was hypothesized that Raman spectroscopy could be used to monitor certain structural and compositional changes associated with such ageing effects and that a relationship between these features and traditional morphology (as measured using a morphology index) could be observed. The hypothesis was tested by measuring spectral features associated with hemoglobin oxygenation from dry-fixed smears and liquid RBCs for twenty-nine different donors (combined), and comparing the trends with morphological scoring from seven of these donors. After appropriately fitting the two data sets to either power or linear curves, the oxygenation state was shown to change in a manner that was donor-dependent and that closely tracked morphological changes. This study suggests Raman analysis has promise for providing a rapid and objective measure of the cell quality of stored RBCs through measurements of hemoglobin oxygenation that is comparable to traditional morphological assessment.

Theoretical probability distributions are often fitted to the individual peaks of Raman spectra to decompose them and facilitate further analyses. Fitting has the additional advantage of eliminating noise. We have exploited this noise-eliminating attribute of fitting procedures in an automated algorithm to smooth Raman spectra. An initial smoothing was performed by fitting Voigt distributions to every channel in a spectrum. The Voigt distribution characters used were strongly Gaussian, the distribution widths equal to the spectral resolution, and their initial amplitudes equal to the spectral intensities at the channels where they were located. The smoothed spectrum was then subtracted from the original noisy spectrum to obtain a residual. For channels where the residual exceeded a limit-of-detection threshold, the distribution width was decreased. The fitting was then repeated until a secondary, lower limit distribution width was reached. The residual was then smoothed repeatedly in the same manner until the minimum distribution width was reached. After each repetition, the smoothed residual was added to the smoothed spectrum. The process was continued until a combined limit of detection and chi-squared stopping criterion was reached. Although slower in comparison to spline- and Savitzky-Golay-based methods, the smoothing quality was significantly better allowing the majority of smoothed spectra, in contrast to these methods, to pass a stringent smoothing quality test.

Individual units of donated red blood cells (RBCs) do not ordinarily undergo analytical testing prior to transfusion. This study establishes the utility of Raman spectroscopy for analyzing the biochemistry of stored RBC supernatant and reveals interesting storage-related changes about the accumulation of lactate, a chemical species that may be harmful to certain patients. The data show measurable variations in supernatant composition and demonstrate that some units of donated RBCs accumulate lactate much more readily than others. The spectra also indicate a higher relative concentration of lactate in units collected from male donors than female donors (p = 0.004) and imply that there is a greater degree of variability at later stages of storage in units from older male donors (>45 years). The study proves that Raman analysis has promise for elucidating the relationship between the metabolism of stored RBCs and donor characteristics. It also suggests that there may be benefit in developing a Raman instrument for the rapid non-invasive assessment of blood-bag biochemistry by measuring through plastic over-layers.

Raman spectroscopy has been used extensively in thermal maturation studies of kerogen, but has not been used to examine the maturation of organic cements in agglutinated foraminifera. Here, we use Raman spectroscopy to document the existence of carbonaceous matter and silica in recent and fossil agglutinated foraminifera, and to measure thermal alteration effects in fossil foraminifera. The distribution of carbonaceous matter through the test (shell) walls of agglutinated foraminifera suggests that this carbonaceous material is derived from primary organic cement and not from random contamination. Fossil specimens exhibit three broad stages of maturation: (1) Immature specimens are characterized by moderately strong fluorescence, broad, low intensity Raman peaks (relative to fluorescence), and a tendency for the G-band to occur at lower wave numbers. These attributes are consistent with the presence of amorphous carbonaceous matter and minor organic degradation. (2) Mature samples (oil window) exhibit high fluorescence, increased relative D- and G-band intensities, and a decreased width of the D-band. (3) Postmature samples exhibit low levels of fluorescence and high relative D- and G-band intensities, a tendency for the G-band to be located at higher wave numbers, an increase in the D:G band ratio, and an increase of the relative intensity of the silica peak. This stage is consistent with the presence of highly ordered carbonaceous matter and diagenetic quartz. These findings indicate that Raman spectroscopic analysis of fossil agglutinated foraminifera can be used as a quick and easy tool to assess thermal maturity and estimate optimal temperatures for hydrocarbon generation.

Supplemental Material for Smoothing Raman Spectra with Contiguous Single-Channel Fitting of Voigt Distributions: An Automated, High-Quality Procedure by H. Georg Schulze, Shreyas Rangan, Michael W. Blades, James M. Piret and Robin F. B. Turner in Applied Spectroscopy

Mammalian cells contain various macromolecules that can be investigated non-invasively with Raman spectroscopy. The particular mixture of major macromolecules present in a cell being probed are reflected in the measured Raman spectra. Determining macromolecular identities and estimating their concentrations from these mixture Raman spectra can distinguish cell types and otherwise enable biological research. However, the application of canonical multivariate methods, such as principal component analysis (PCA), to perform spectral unmixing yields mathematical solutions that can be difficult to interpret. Non-negative matrix factorization (NNMF) improves the interpretability of unmixed macromolecular components, but can be difficult to apply because ambiguities produced by overlapping Raman bands permit multiple solutions. Furthermore, theoretically sound methods can be difficult to implement in practice. Here we examined the effects of a number of empirical approaches on the quality of NNMF results. These approaches were evaluated on simulated mammalian cell Raman hyperspectra and the results were used to develop an enhanced procedure for implementing NNMF. We demonstrated the utility of this procedure using a Raman hyperspectral data set measured from human islet cells to recover the spectra of insulin and glucagon. This was compared to the relatively inferior PCA of these data.

Charge-coupled device detectors are vulnerable to cosmic rays that can contaminate Raman spectra with positive going spikes. Because spikes can adversely affect spectral processing and data analyses, they must be removed. Although both hardware-based and software-based spike removal methods exist, they typically require parameter and threshold specification dependent on well-considered user input. Here, we present a fully automated spike removal algorithm that proceeds without requiring user input. It is minimally dependent on sample attributes, and those that are required (e.g., standard deviation of spectral noise) can be determined with other fully automated procedures. At the core of the method is the identification and location of spikes with coincident second derivatives along both the spectral and spatiotemporal dimensions of two-dimensional datasets. The method can be applied to spectra that are relatively inhomogeneous because it provides fairly effective and selective targeting of spikes resulting in minimal distortion of spectra. Relatively effective spike removal obtained with full automation could provide substantial benefits to users where large numbers of spectra must be processed.

The cell cycle is a series of integrated and coordinated physiological events that results in cell growth and replication. Besides observing the event of cell division it is not feasible to determine the cell cycle phase without fatal and/or perturbing invasive procedures such as cell staining, fixing, and/or dissociation. Raman microspectroscopy (RMS) is a chemical imaging technique that exploits molecular vibrations as a contrast mechanism; it can be applied to single living cells noninvasively to allow unperturbed analysis over time. We used RMS to determine the cell cycle phase based on integrating the composite 783 cm(-1) nucleic acid band intensities across individual cell nuclei. After correcting for RNA contributions using the RNA 811 cm(-1) band, the measured intensities essentially reflected DNA content. When quantifying Raman images from single cells in a population of methanol-fixed human embryonic stem cells, the histogram of corrected 783 cm(-1) band intensities exhibited a profile analogous to that obtained using flow-cytometry with nuclear stains. The two population peaks in the histogram occur at Raman intensities corresponding to a 1-fold and 2-fold diploid DNA complement per cell, consistent with a distribution of cells with a population peak due to cells at the end of G1 phase (1-fold) and a peak due to cells entering M phase (2-fold). When treated with EdU to label the replicating DNA and block cell division, cells with higher EdU-related fluorescence generally had higher integrated Raman intensities. This provides proof-of-principle of an analytical method for label-free RMS determination in situ of cell cycle phase in adherent monolayers or even single adherent cells.

Raman spectra often contain undesirable, randomly positioned, intense, narrow-bandwidth, positive, unidirectional spectral features generated when cosmic rays strike charge-coupled device cameras. These must be removed prior to analysis, but doing so manually is not feasible for large data sets. We developed a quick, simple, effective, semi-automated procedure to remove cosmic ray spikes from spectral data sets that contain large numbers of relatively homogenous spectra. Although some inhomogeneous spectral data sets can be accommodated—it requires replacing excessively modified spectra with the originals and removing their spikes with a median filter instead—caution is advised when processing such data sets. In addition, the technique is suitable for interpolating missing spectra or replacing aberrant spectra with good spectral estimates. The method is applied to baseline-flattened spectra and relies on fitting a third-order (or higher) polynomial through all the spectra at every wavenumber. Pixel intensities in excess of a threshold of 3 × the noise standard deviation above the fit are reduced to the threshold level. Because only two parameters (with readily specified default values) might require further adjustment, the method is easily implemented for semi-automated processing of large spectral sets.

A fully automated and model-free baseline-correction method for vibrational spectra is presented. It iteratively applies a small, but increasing, moving average window in conjunction with peak stripping to estimate spectral baselines. Peak stripping causes the area stripped from the spectrum to initially increase and then diminish as peak stripping proceeds to completion; a subsequent increase is generally indicative of the commencement of baseline stripping. Consequently, this local minimum is used as a stopping criterion. A backup is provided by a second stopping criterion based on the area under a third-order polynomial fitted to the first derivative of the current estimate of the baseline-free spectrum and also indicates whether baseline is being stripped. When the second stopping criterion is triggered instead of the first one, a proportionally scaled simulated Gaussian baseline is added to the current estimate of the baseline-free spectrum to act as an internal standard to facilitate subsequent processing and termination via the first stopping criterion. The method is conceptually simple, easy to implement, and fully automated. Good and consistent results were obtained on simulated and real Raman spectra, making it suitable for the fully automated baseline correction of large numbers of spectra.

Human embryonic stem cells (hESCs) have large nucleus-to-cytoplasm ratios and nucleic acid spectral bands are prominent in their characteristic Raman signatures. Under normal conditions, the major variations in these signatures are due to changes in glycogen content, but how these signatures vary in response to different external conditions is largely unknown. In this study we investigated the influences of temperature variations on hESC Raman signatures. At 32 °C, compared to the 37 °C control condition, cell proliferation rates were markedly reduced and glycogen Raman band intensities were elevated. In addition, at both temperatures, an inverse relationship between cell proliferation rates (i.e., onset of exponential growth phase vs. end of exponential phase) and glycogen Raman band intensities was observed. This relationship suggested a role for glycogen in the energy metabolism of hESC self-renewal. Protein and lipid spectral variations were small and co-varied with those of nucleic acids, suggesting that they were related to changes in cellular dimensions occurring during the cell cycle. When the temperature was elevated to 39 °C, increased glycogen band intensities, compared to controls, were also observed. In addition, spectral evidence of differentiation emerged that was supported by reduced SSEA-3 expression. Taken together, these results demonstrated that heat and cold stress had quite different effects on the characteristic Raman signatures of hESCs. Thus, Raman spectroscopy can be used to detect deviation from optimal culturing temperatures and therefore it could be of considerable value in the routine and noninvasive determination of hESC culture quality.

Raman microspectroscopy is a quantitative instrumental method with considerable promise for the nondestructive analysis of living biological samples. Amongst samples of particular interest are human embryonic stem cells because of their therapeutic potential and because examination using Raman microspectroscopy does not appear to adversely affect this potential. However, it can be difficult to compare different spectra obtained with this technique and to quantify the native cellular constituents of such samples because their characteristic dimensions are difficult to establish or may vary from point to point. We present here a method to normalize spectra and estimate sample thicknesses based on a reference component present in the basal cell culture medium when we perform spectroscopy on colonies of living cells. Because more basal medium is displaced from the sampling volume as the cell layer increases in thickness, and because this component is present in the medium but excluded from cells, a concomitant decline therefore occurs in the intensity of the Raman scattering from the reference component. This permits comparisons between samples because their spectra can be scaled in inverse relation to their excluded volumes. Furthermore, estimations of sample thicknesses can also be obtained based on the same concept. Thus, the absolute quantification of cellular components becomes possible because cell sample volumes can be determined. Although applied to human embryonic stem cells, the approach is sufficiently general to be adapted for use with other samples. Copyright © 2011 John Wiley & Sons, Ltd.

We present a method to perform absolute quantification of glycogen in human embryonic stem cells (hESCs) in situ based on the use of Raman microspectroscopy. The proposed quantification method was validated by comparison to a commonly used commercial glycogen assay kit. With Raman microspectroscopy, we could obtain the glycogen content of hESCs faster and apparently more accurately than with the kit. In addition, glycogen distributions across a colony could be obtained. Raman spectroscopy can provide reliable estimates of the in situ glycogen content in hESCs, and this approach should also be extensible to their other biochemical constituents as well as to other cell types.

Spectroscopic methods based on Raman scattering have for many years employed synthetic oligonucleotides in a broad range of applications, either as probes or as model analytes for biophysical investigations. Benzamide is commonly used as a protecting group in the phosphoramidite synthesis of oligonucleotides and, while standard desalting used after synthesis yields sufficiently pure reagents for most assay reactions or other routine uses of the oligomer, it does not completely remove benzamide. We show that the 1609 cm−1 band of residual benzamide contamination can interfere with certain nucleic acid bands, particularly when using excitation wavelengths near 244 nm where the benzamide band is strongly enhanced. For example, the 1609 cm−1 band of benzamide could obscure (or be mistaken as) a weak vibration attributed to an NH2 scissoring. The extent of benzamide contamination in desalted preparations varies considerably among different commercial sources, and hence caution is advised when making direct comparisons of ultraviolet Raman data of oligonucleotides from different sources. Copyright © 2010 John Wiley & Sons, Ltd.

We present here a fully automated spectral baseline-removal procedure. The method uses a large-window moving average to estimate the baseline; thus, it is a model-free approach with a peak-stripping method to remove spectral peaks. After processing, the baseline-corrected spectrum should yield a flat baseline and this endpoint can be verified with the χ2-statistic. The approach provides for multiple passes or iterations, based on a given χ2-statistic for convergence. If the baseline is acceptably flat given the χ2-statistic after the first pass at correction, the problem is solved. If not, the non-flat baseline (i.e., after the first effort or first pass at correction) should provide an indication of where the first pass caused too much or too little baseline to be subtracted. The second pass thus permits one to compensate for the errors incurred on the first pass. Thus, one can use a very large window so as to avoid affecting spectral peaks—even if the window is so large that the baseline is inaccurately removed—because baseline-correction errors can be assessed and compensated for on subsequent passes. We start with the largest possible window and gradually reduce it until acceptable baseline correction based on the χ2-statistic is achieved. Results, obtained on both simulated and measured Raman data, are presented and discussed.

Human embryonic stem cells (hESCs) have typical Raman signatures, but specific factors that contribute to variations in these signatures have not been reported to date. Furthermore, variations due to the passaging that is necessary for hESC culture maintenance could potentially distort these signatures. It is therefore important to characterize the impact of these culture manipulations on the Raman spectra to gain a better understanding of the origins and nature of their variations. Here we report on the Raman microspectroscopy of hESCs samples from maintenance cultures, complemented with periodic acid Schiff (PAS, carbohydrates) and 4′-6-diamidino-2-phenlyindol (DAPI, nuclei) staining. The component predominantly responsible for variations between spectra was spectrally identified as glycogen. Variations in the Raman map of the 480 cm−1 glycogen marker band corresponded with those of a PAS stain of the same sample area. The 785-nm Raman microspectra of hESC cultures examined daily after passaging showed that the same nonrandom spectral variances occurred at all time points after passaging. The pattern of these variances was identified as being due to glycogen spectral components. Our results help validate the previously observed spectral signatures of hESCs and further delineate and characterize the variations that can be expected in these signatures under normal maintenance culture conditions, and aid distinguishing them from those corresponding to differentiation, thus providing a benchmark for future studies. Copyright © 2010 John Wiley & Sons, Ltd.

We present a new spectral image processing algorithm, the “matrix maximum entropy method” (MxMEM), which offers efficient signal-to-noise ratio (SNR) enhancement of multidimensional spectral data. MxMEM is based upon two previous regularization methods that employ the maximum entropy concept. The first is a one-dimensional (1D) algorithm, which smoothes individual vectors, called the two-point maximum entropy method (TPMEM). The second is a two-dimensional (2D) form called 2D TPMEM, that smoothes images but processes them one vector at a time. MxMEM is a truly two dimensional image processing algorithm in that its “smoothing engine” performs two-dimensional processing in every iteration. We demonstrate that this matrix-based construction makes more effective use of two-dimensionally embedded information and thus confers significant advantages over other regularization approaches. In addition, we utilize the concept that individual related Raman spectra can be combined in a matrix to form an artificial Raman “image”. We show that, when processed as an image, superior SNR enhancement is achieved compared to processing the same data by TPMEM one spectrum at a time.

Raman microspectroscopy allows the classification of populations of human embryonic stem cells (hESCs) in different stages of differentiation based on the relative intensities of certain amino acid and nucleic acid bands. Here, we report the results of a comparative study of the Raman spectra of live cells versus cells killed and fixed by rapid desiccation, focusing on the ratio of intensities at 757 cm−1 (tryptophan) and 784 cm−1 (DNA and RNA). We observe that the same temporal pattern emerges over a 3-week time course in both sample types. This suggests that prolonged observations of dry-fixed cells can yield high signal-to-noise chemical images that cannot be obtained from colonies of living cells where the time scale of significant biological changes are comparable to the time scale of the measurement. This permits, for example, comparison of the spatial distributions of cells at different stages of differentiation within the same colonies. Copyright © 2010 John Wiley & Sons, Ltd.

Raman microspectroscopy is an attractive approach for chemical imaging of biological specimens, including live cells, without the need for chemi-selective stains. Using a microspectrometer, near-infrared Raman spectra throughout the range 663 cm(-1) to 1220 cm(-1) were obtained from colonies of CA1 human embryonic stem cells (hESCs) and CA1 cells that had been stimulated to differentiate for 3 weeks by 10% fetal bovine serum on gelatin. Distributions and intensities of spectral bands attributed to proteins varied significantly between undifferentiated and differentiated cells. Importantly, compared to proteins and lipids, the band intensities of nucleic acids were dominant in undifferentiated cells with a dominance-reversal in differentiated cells. Thus, we could identify intensity ratios of particular protein-related bands (e.g., 757 cm(-1) tryptophan) to nucleic acid bands (784 cm(-1) DNA/RNA composite) that were effective in discriminating between spectra of undifferentiated and differentiated cells. We observed no discernible negative effects due to the laser exposure in terms of morphology, proliferation, or pluripotency of the stem cells. We conclude that Raman microscopy and complementary data processing procedures provide a rapid, noninvasive approach that can distinguish hESCs from differentiated cells. This is the first report to identify specific Raman markers for the differentiation status of hESCs.

We report here the first UV resonance Raman spectroscopic (UVRRS) study on locked nucleic acid (LNA) oligomers. Locking a base in nucleic acid (NA) oligomers produces a conformational change in the glycosyl bond between backbone and base. We present evidence of this change in LNAs when compared to their natural analogs using UVRRS. Wavenumber downshifts and peak amplitude increases, especially of the ∼1481 cm−1 peak that is a spectral marker for part of the glycosyl bond, correlate with the fraction of locked bases when single-stranded oligomers incorporating up to three locked bases were examined. By varying the position of the locked base within a fixed length sequence, we conclude that one, or at most two bases, on either side of the lock is affected. We further conclude from these data, and previously published reports, that the conformation of LNA is determined by imidazole–imidazole and pyrimidine–pyrimidine repulsion and imidazole–pyrimidine attraction in contrast to dispersion attraction-dependent aggregation in the B conformation of DNA. Copyright © 2009 John Wiley & Sons, Ltd.

Transient absorption measurements of excited states in DNA and LNA were performed using a femtosecond pump-probe arrangement with excitation at 266 nm and absorption monitored at 400 nm while varying the sample tem- perature between 5 °C and 70 °C. Samples consisted of adenine monophosphate monomer, polyadenine 12-mer in single- stranded form, and polyadenine 12-mer in hybridized form. Excited states decayed in a biphasic manner with short-lived ( 1) and long-lived ( 2) components, while the monomer had only a 'single' short-lived decay time. Temperature increases increased absorption intensities and reduced  1 until they approached those of the monomer at high temperatures (where stacking is minimal). These results suggest that the initial excitation in stacked regions is cooperative and involves several bases and that the number of bases involved is reduced with increasing temperature. In contrast, increasing temperatures had little effect on  2 while absorption intensities decreased, suggesting that very few, perhaps only two, stacked bases are involved and that their number is reduced at higher temperatures. We found no clear evidence of melting point transitions indicating that those excited states probed with our arrangement were not dependent on base pairing. Our results are con- sistent with and strengthen an emerging consensus model of excited state dynamics in DNA wherein a UV photon is ab- sorbed collectively by electronically coupled and thus well-stacked intrachain bases. This collective excitation results in a Frenkel exciton that is delocalized over these bases, and the Frenkel exciton then decays rapidly to a long-lived, lower en- ergy, dark intrachain exciplex.

The two-point maximum entropy method (TPMEM) is a useful method for signal-to-noise ratio enhancement and deconvolution of spectra, but its efficacy is limited under conditions of high background offsets. This means that spectra with high average background levels, regions with high background in spectra with varying background levels, and regions of high signal-to-noise ratios are smoothed less effectively than spectra or spectral regions without these conditions. We report here on the cause of this TPMEM limitation and on appropriate baseline estimation and removal procedures that effectively minimize the effects on regularization. We also present a comparative analysis of TPMEM and Savitzky–Golay filtering to facilitate selection of the best technique under a given range of conditions.

Continued advances toward cell-based therapies for human disease generate a growing need for unbiased and label-free monitoring of cellular characteristics. We used Raman microspectroscopy to characterize four important stages in the 26-day directed differentiation of human embryonic stem cells (hESCs) to insulin-positive cells. The extent to which the cells retained spectroscopic features of pluripotent cells or developed spectroscopic features suggestive of pancreatic endocrine cells, as well as assessing the homogeneity of the cell populations at these developmental stages, were of particular interest. Such information could have implications for the utility of Raman microspectroscopy process analysis for the generation of insulin-positive cells from hESCs. Because hESC seeding density influences the subsequent pancreatic development, three different seeding density cultures were analyzed. Transcription factor and other marker analyses assessed the progress of the cells through the relevant developmental stages. Increases in the Raman protein-to-nucleic acid band ratios were observed at the final endocrine stage analyzed, but this increase was less than expected. Also, high glycogen band intensities, somewhat unexpected in pancreatic endocrine cells, suggested the presence of a substantial number of glycogen containing cells. We discuss the potential process analytical technology application of these findings and their importance for cell manufacturing.

Modern transfusion medicine relies on the safe, secure, and cost-effective delivery of donated red blood cells (RBCs). Once isolated, RBCs are suspended in a defined additive solution and stored in plastic blood bags in which, over time, they undergo chemical, physiological, and morphological changes that may have a deleterious impact on some patients. Regulations limit the storage period to 42 days and the cells do not routinely undergo analytical testing before use. In this study, we use Raman spectroscopy to interrogate stored RBCs and we identify metabolic and cell-breakdown products, such as haemoglobin and membrane fragments, that build-up in the blood bags as the cells age. Our work points the way to the development of an instrument which could quickly and easily assess the biochemical nature of stored RBC units before they are transfused.

High-throughput information extraction from large numbers of Raman spectra is becoming an increasingly taxing problem due to the proliferation of new applications enabled using advances in instrumentation. Fortunately, in many of these applications, the entire process can be automated, yielding reproducibly good results with significant time and cost savings. Information extraction consists of two stages, preprocessing and analysis. We focus here on the preprocessing stage, which typically involves several steps, such as calibration, background subtraction, baseline flattening, artifact removal, smoothing, and so on, before the resulting spectra can be further analyzed. Because the results of some of these steps can affect the performance of subsequent ones, attention must be given to the sequencing of steps, the compatibility of these sequences, and the propensity of each step to generate spectral distortions. We outline here important considerations to effect full automation of Raman spectral preprocessing: what is considered full automation; putative general principles to effect full automation; the proper sequencing of processing and analysis steps; conflicts and circularities arising from sequencing; and the need for, and approaches to, preprocessing quality control. These considerations are discussed and illustrated with biological and biomedical examples reflecting both successful and faulty preprocessing.

Ultraviolet resonance Raman spectroscopy (UVRRS) was used to monitor UV radiation and thermal energy deposition in 12- and 18-mer AT oligomers. Difference spectroscopy and two-dimensional correlation spectroscopy (2DCOS) were used to characterize the distinct spectral responses manifested by the two processes, and to examine their potential cooperative and/or independent effects on the ensemble spectrum. Experiments utilizing incremental doses of 257-nm radiation revealed that the most affected bands are those involving double bonds when the sample temperature was held constant. Complementary experiments utilizing sample heating indicated that the most affected bands were those involving hydrogen-bond disruption. Finally, it is shown that bands associated with Watson–Crick hydrogen-bond disruption are the most affected in rapidly heat-cycled samples. The results also suggest that both radiation and thermal effects produce independent structural changes in AT oligomers and, furthermore, that these effects are complementary when the experiments involve single-strand DNA fragments. When using UVRRS to study one of these perturbations (e.g. thermal stability of DNA), the concurrent perturbation (e.g. UV exposure) must not be neglected, since UV exposure is inherent in the UVRR process. These findings thus have implications for the use of UVRRS in the study of DNA dynamics. Copyright © 2006 John Wiley & Sons, Ltd.

Rapid technological advances have made the acquisition of large numbers of spectra not only feasible, but also routine. As a result, a significant research effort is focused on semi-automated and fully automated spectral processing techniques. However, the need to provide initial estimates of the number of peaks, their band shapes, and the initial parameters of these bands presents an obstacle to the full automation of peak fitting and its incorporation into fully automated spectral-preprocessing workflows. Moreover, the sensitivity of peak-fit routines to initial parameter settings and the resultant variations in solution quality further impede user-free operation. We have developed a technique to perform fully automated peak fitting on fully automated preconditioned spectra—specifically, baseline-corrected and smoothed spectra that are free of cosmic-ray-induced spikes. Briefly, the tallest peak in a spectrum is located and a Gaussian peak-fit is performed. The fitted peak is then subtracted from the spectrum, and the procedure is repeated until the entire spectrum has been processed. In second and third passes, all the peaks in the spectrum are fitted concurrently, but are fitted to a Pearson Type VII model using the parameters for the model established in the prior pass. The technique is applied to a synthetic spectrum with several peaks, some of which have substantial overlap, to test the ability of the method to recover the correct number of peaks, their true shape, and their appropriate parameters. Finally the method is tested on measured Raman spectra collected from human embryonic stem cells and samples of red blood cells.

Crystalline silicon, widely used in the electronic industry, is also a very popular material for calibrating Raman spectrometry instruments. Silicon chips cut or cleaved from commercially available silicon wafers are low-cost monolithic monocrystalline materials that give a strong Raman line at 521 cm(-1) with almost no background. Such chips have at least one optically flat surface and can be used in place of glass microscope slides as sample substrates that provide an internal calibration standard during Raman measurements. The Raman signal intensity from the silicon can be selectively attenuated by depositing a gold layer on top of the silicon surface with variable thickness such that the far-field silicon Raman signal is comparable with the Raman signal of an investigated material adjacent to this structure. This gold layer provides the additional advantage of increased sensitivity of the spectral signal from the sample due to the reflectivity of the gold surface, which allows forward and backscattered analyte Raman excitation and signal collection. An additional thin encapsulating overlayer of SiO2 provides a protective and biocompatible surface to facilitate Raman microspectroscopic investigation of live cells.

The nucleolus is a prominent subnuclear structure whose major function is the transcription and assembly of ribosome subunits. The size of the nucleolus varies with the cell cycle, proliferation rate and stress. Changes in nucleolar size, number, chemical composition, and shape can be used to characterize malignant cells. We used spontaneous Raman microscopy as a label-free technique to examine nucleolar spatial and chemical features. Raman images of the 1003 cm(-1) phenylalanine band revealed large, well-defined subnuclear protein structures in MFC-7 breast cancer cells. The 783 cm(-1) images showed that nucleic acids were similarly distributed, but varied more in intensity, forming observable high-intensity regions. High subnuclear RNA concentrations were observed within some of these regions as shown by 809 cm(-1) Raman band images. Principal component analyses of sub-images and library spectra validated the subnuclear presence of RNA. They also revealed that an actin-like protein covaried with DNA within the nucleolus, a combination that accounted for 64% or more of the spectral variance. Embryonic stem cells are another rapidly proliferating cell type, but their nucleoli were not as large or well defined. Estimating the size of the larger MCF-7 nucleolus was used to show the utility of Raman microscopy for morphometric analyses. It was concluded that imaging based on Raman microscopy provides a promising new method for the study of nucleolar function and organization, in the evaluation of drug and experimental effects on the nucleolus, and in clinical diagnostics and prognostics.

In this paper, a model of lordosis in the female rat is proposed, based on possible neurochemical events mediating this behavior. In particular, we hypothesize that the synthesis of a specific neurotransmitter may be the rate-limiting step in the onset of lordosis behavior. Based on this hypothesis, the display of lordosis reflects the availability of this neurotransmitter. We postulate several alternative chemical reaction mechanisms to account for the synthesis of this transmitter. Behavioral testing, using ovariectomized estrogen-primed female rats, during the period immediately following progesterone injection, has enabled us to select one of these mechanisms. The implications of our model, offering a new method for studying mechanisms of behavior, are discussed.

A new method for Raman signal recovery, the two-point maximum entropy method (TPMEM), based on a regularization method using two-point entropy is presented. The method can be used for signal-to-noise ratio (SNR) enhancement in very low SNR measurements or for deconvolution, in order to remove the effects of the instrumental line shape on the measured spectrum. Unlike most SNR enhancement schemes, TPMEM requires no filter parameters and no a priori knowledge of the expected signal. A rigorous test on a randomly produced set of convolved and/or noise-corrupted simulated Raman spectra is presented in order to validate the method and compare it to Savitzky-Golay filtering and the maximum entropy method. The method is evaluated on the basis of the root mean square (rms) error and correlation coefficients of the recovered data with the original data, as well as on the basis of SNR improvement, and showed significant improvements in both performance and speed over conventional methods. The method is demonstrated in an application involving fiber-optic-linked Raman and resonance Raman spectroscopy.

Identification of individual components in biological mixtures can be a difficult problem regardiess of the analytical method employed. In this work, Raman spectroscopy was chosen as a prototype analytical method due to its inherent versatility and applicability to aqueous media, making it useful for the study of biological samples. Artificial neural networks (ANNs) and the classical least-squares (CLS) method were used to identify and quantify the Raman spectra of the small-molecule neurotransmitters and mixtures of such molecules. The transfer functions used by a network, as well as the architecture of a network, played an important role in the ability of the network to identify the Raman spectra of individual neurotransmitters and the Raman spectra of neurotransmitter mixtures. Specifically, networks using sigmoid and hyperbolic tangent transfer functions generalized better from the mixtures in the training data set to those in the testing data sets than networks using sine functions. Networks with connections that permit the local processing of inputs generally performed better than other networks on all the testing data sets, and better than the CLS method of curve fitting, on novel spectra of some neurotransmitters. The CLS method was found to perform well on noisy, shifted, and difference spectra.

We have shown that neural networks are capable of accurately identifying the Raman spectra of aqueous solutions of small-molecule neurotransmitters. It was found that the networks performed optimally when the ratio of the number of hidden nodes to the number of input nodes was 0.16, that network accuracy increased with the number of input layer nodes, and that input features influenced the abilities of networks to discriminate or generalize between spectra. Furthermore, networks employing sine transfer functions for their hidden layers trained faster and were better at discriminating between closely related spectra, but they were less tolerant of spectral distortions than the networks using sigmoid transfer functions. The latter type of network produced superior results where generalization between spectra was required.

Measuring spatial and temporal patterns of cytochemical variation in human embryonic stem cell (hESC) colonies is necessary for understanding the role of cellular communication in spontaneous differentiation, the mechanisms of biological niche creation, and structure-generating developmental processes. Such insights will ultimately facilitate directed differentiation and therewith promote advances in tissue engineering and regenerative medicine. However, the patterns of cytochemical inhomogeneities of hESC colonies are not well studied and their causes are not fully understood. We used Raman spectroscopic mapping to contrast supracellular variations in cytochemical composition across pluripotent and partly differentiated hESC colonies to gain a better understanding of the early-stage (i.e., 5 days) effects of the differentiation process on the nature and evolution of these patterns. Higher protein-to-nucleic acid ratios, a differentiation status indicator observed previously using Raman spectroscopy, confirmed reported results that spontaneous differentiation is more pronounced on the edges of a colony than elsewhere. In addition, pluripotent and partly differentiated colonies also showed higher lipid concentrations relative to nucleic acids at colony edges in contrast to relative glycogen concentrations, which were up to 400% more pronounced in the colony centers compared to their edges. Pluripotent and partly differentiated colonies differed, with the latter having higher average protein-to-nucleic acid and lipid-to-nucleic acid ratios but a lower glycogen-to-nucleic acid ratio. In both cases, cell density, pluripotency, and high glycogen appeared to vary in tandem. Spatial variations in glycogen- and protein-to-nucleic acid ratios have features on the order of 100 l m and larger. These dimensions are consistent with those reported for stem cell niches and suggest that cytochemical inhomogeneities may provide colony-level information about niches and niche formation. These results demonstrate Raman mapping to be a potentially useful technique for revealing the complexities in the spatial organization of hESC cultures and thus for monitoring the evolution of engineered hESC niches.

Two-dimensional correlation spectroscopy (2D-COS) is a powerful spectral analysis technique widely used in many fields of spectroscopy because it can reveal spectral information in complex systems that is not readily evident in the original spectral data alone. However, noise may severely distort the information and thus limit the technique's usefulness. Consequently, noise reduction is often performed before implementing 2D-COS. In general, this is implemented using one-dimensional (1D) methods applied to the individual input spectra, but, because 2D-COS is based on sets of successive spectra and produces 2D outputs, there is also scope for the utilization of 2D noise-reduction methods. Furthermore, 2D noise reduction can be applied either to the original set of spectra before performing 2D-COS ("pretreatment") or on the 2D-COS output ("post-treatment"). Very little work has been done on post-treatment; hence, the relative advantages of these two approaches are unclear. In this work we compare the noise-reduction performance on 2D-COS of pretreatment and post-treatment using 1D (wavelets) and 2D algorithms (wavelets, matrix maximum entropy). The 2D methods generally outperformed the 1D method in pretreatment noise reduction. 2D post-treatment in some cases was superior to pretreatment and, unexpectedly, also provided correlation coefficient maps that were similar to 2D correlation spectroscopy maps but with apparent better contrast.

The automated processing of data from high-throughput and real-time collection procedures is becoming a pressing problem. Currently the focus is shifting to automated smoothing techniques where, unlike background subtraction techniques, very few methods exist. We have developed a filter based on the widely used and conceptually simple moving average method or zero-order Savitzky–Golay filter and its iterative relative, the Kolmogorov–Zurbenko filter. A crucial difference, however, between these filters and our implementation is that our fully automated smoothing filter requires no parameter specification or parameter optimization. Results are comparable to, or better than, Savitzky–Golay filters with optimized parameters and superior to the automated iterative median filter. Our approach, because it is based on the highly familiar moving average concept, is intuitive, fast, and straightforward to implement and should therefore be of immediate and considerable practical use in a wide variety of spectroscopy applications.

The ability of ultraviolet resonance Raman spectroscopy (UVRRS) to determine structural, environmental, and analytical information concerning low-concentration aqueous biomolecules makes it a powerful bioanalytical and biophysical technique. Unfortunately, its utility has been limited by experimental requirements that preclude in situ or in vivo studies in most cases. We have developed the first high-performance fiber-optic probes suitable for long-term use in pulsed UVRRS applications in the deep- UV (DUV, 205-250 nm). The probes incorporate recently developed improved ultraviolet (IUV) fibers that do not exhibit the rapid solarization and throughput decay that previously hampered the use of optical fibers for delivering pulsed, DUV light. A novel 90 degrees mirrored collection geometry is used to overcome the inner-filtering effects that plague flush-probe geometries. The IUV fibers are characterized with respect to their efficacy at transmitting pulsed, DUV laser light, and prototype probes are used to obtain pulsed UVRRS data of aromatic amino acids, proteins, and hormones at low concentrations with 205-240-nm pulsed excitation. Efficient probe geometries and fabrication methods are presented. The performance of the probes in examining resonance-enhanced Raman signals from absorbing chromophores is investigated, and the optimal excitation wavelength is shown to be significantly red-shifted from the maximum of the resonance Raman enhancement profile. Generally applicable procedures for determining optimal experimental conditions are also introduced.

Embryonic stem cells (ESC), derived from preimplantation embryos, are defined by their ability to both self-renew and differentiate into all of the cells and tissues of a mature animal. Efforts to develop methods for in vitro culture of ESC for research or eventual therapeutic applications are hampered by the lack of rapid, nondestructive assays for distinguishing ESC from other (differentiated) cells within a growing culture. Coherent anti-Stokes Raman scattering (CARS) microscopy is shown here to be a sensitive and nondestructive method for identifying mouse ESC based on selective observation of specific molecular vibrations believed to be spectroscopic markers indicating the differentiated vs undifferentiated states of such cells. The nonlinear nature of CARS also permits imaging with subcellular resolution, potentially offering a means by which chemical changes accompanying the early stages of differentiation may be associated with certain intracellular compartments (e.g., nucleus, cytoplasm, membranes). A novel exposure/collection configuration is described, which yields high collection efficiency and low interference from nonresonant background components.

The above-ground organs of plants are covered by a cuticle, an extracellular membrane performing important physiological and ecological functions, that consists of the fatty acid-derived polymer cutin and waxes. In the cuticular wax of many species, including the leaves of Prunus laurocerasus, triterpenoids are found at high concentrations. This paper investigates the potential of Raman microspectroscopy for the simultaneous detection of structurally similar triterpenoids in plant cuticles. Relative composition analysis was first performed on artificial triterpenoid mixtures consisting of α-amyrin and oleanolic acid, as well as oleanolic acid and ursolic acid, the two triterpenoids abundantly found in the cuticles of P. laurocerasus. The different triterpenoids could be distinguished in the mixture spectra and the resulting calculated triterpenoid ratios were consistent with the expected values. Qualitative analysis of the Raman spectra of P. laurocerasus cuticle demonstrated the in situ detectability of the triterpenoids using this approach. It is shown here that Raman microspectroscopy has the potential to provide useful information concerning the spatial distribution of some key chemical components of plant cuticles. This technique thus offers a valuable complement to the current standard analytical methods used for analyzing the bulk composition of plant cuticles.