James Piret and Robin Turner of Michael Smith Laboratories (Vancouver, BC, Canada) and the University of British Columbia (UBC) have been exploring the benefits of extracting and displaying spectrometric and non-spectrometric variables correlated with a proposed method called multifont correlation. . analysis (MuSCA). His work has discovered several advantages of using Raman spectroscopy for these applications. Here, they discuss their efforts to develop an approach that would allow the integration of diverse biochemical information with measured spectra for coanalysis to characterize the spectra and leverage the available spectral information.
Piret and Turner are recipients of the William F. Meggers Award from the Society for Applied Spectroscopy (SAS) 2022, which is presented annually at FACSS SciX. This award is given to the author (s) of the outstanding work appearing in Applied Spectroscopy. More details about the award are available from the Society for Applied Spectroscopy
In a recent article that was selected for the William F. Meggers 2022 Award from the Society for Applied Spectroscopy (1), you examined the benefits of extracting and displaying correlated spectrometric and non-spectrometric variables. Can you briefly describe this concept, the benefits, and how this method compares to the well-known two-dimensional correlation spectroscopy (2D-COS)?
Conventional 2D-COS analysis can be very useful for assessing correlated and anti-correlated changes in different parts of a spectrum based on some system perturbation being investigated. Therefore, it can provide useful information on how the system responds to this disturbance. However, the interpretation of 2D-COS data can be quite difficult in the case of chemically complex systems such as living cells for which spectra have superimposed characteristics that can have multi-component contributions on all wave numbers. . For example, nucleic acids have several vibratory bands in the region of fingerprints and some overlap with protein bands, some with lipid bands, some with carbohydrate bands, and many overlap with abundant small molecules that have chemical fragments in them. common with nucleic acids.
Independent (non-spectroscopic) measurements that provide relevant monovariant data can often aid in the interpretation of spectral data by providing a direct measurement of some components represented in the spectra. For example, biochemical assays of total proteins or total nucleic acids can help measure the amount of signal superimposed on the bands where both contribute. Likewise, highly selective assays of specific proteins or nucleic acids can help interpret spectral data. In general, any independent measurement can aid in interpretation and link biological variables to observed changes in overlapping spectral peaks. However, these inferred links can usually only be made between the measured data and the characteristics of the spectra that are known to correspond to the measurand, typically specific peaks. It would be even more useful if correlated spectral characteristics that could be distributed across the spectrum could be identified and evaluated. This would be feasible if non-spectrometric data could be analyzed in conjunction with spectral data, perhaps in a properly modified form of 2D-COS analysis.
Our main author, Dr. Georg Schulze, proposed a method he called “multifunctional correlation analysis” (MuSCA) by which any available non-spectroscopic data from the same sample could be encoded along with the measured spectra for each point of perturbation. Non-spectrometric data are encoded by uniformly spaced artificial peaks with uniform widths and with amplitudes scaled appropriately in relation to the value of the non-spectrometric amplitudes (e.g., biochemical assay) at each point of perturbation. These artificial peaks are added to the spectra measured at each point of perturbation. The resulting set of augmented hyperspectral data using 2D-COS is then analyzed to obtain maps (such as covariance maps and correlation coefficient) that directly provide correlations between spectral and non-spectral data across the spectrum. It is essentially an adaptation or extension of the concept used in the perturbation domain (PDD) decomposition.
What inspired you to work on this method and approach? Did the methods currently used need to be improved?
We are exploring the ability of Raman spectroscopy to monitor cell changes during the manufacture of therapeutic cells. Cellular changes during these processes are usually manifested by changes in their biochemical composition that can be very informative regarding the state and quality of cells (e.g., increased or decreased levels of specific proteins). , as well as other changes in macromolecular composition). These changes can be measured directly by biochemical assays, but these measurements typically depend on costly and destructive assays that require samples to be removed from the cells of the system. Frequent sampling can significantly decrease the number of valuable cells to treat patients or even cause microbial contamination, putting at risk the ability to perform life-saving therapies. In addition, delaying time before cell assay results become available may compromise the ability to respond effectively. Therefore, there is a considerable incentive to reduce the need for offline sampling and analysis.
Spectroscopic measurements offer the potential for non-destructive in situ measurements that obviate the need to sample and return results much more quickly (even almost in real time). Infrared (IR) and Raman spectroscopy have been shown to provide data-rich data on cells and tissues, but the data can be difficult to interpret as indicated above. This is especially true if biochemical measurements do not have a clear relationship to spectral data. For example, measurements of transcription factors in cells may indicate the expression of genes important for development. These measurements may correlate with certain characteristics of the measured spectra, although it may not be clear a priori which spectral characteristics are relevant or why.
Multivariate analytical methods such as principal component analysis (PCA) can often reveal qualitative differences in spectral data that are useful, but it is usually difficult to infer the biological origin of the separation observed in PCs. We were drawn to the prospect of using analytical approaches such as 2D-COS that provide quantifiable statistics related to spectral variations. We hoped to be able to characterize the spectroscopic data well enough during the development of the method so that the sampling of the product during manufacturing could be significantly reduced. Thus, an attempt was made to develop an approach that would allow the integration of diverse biochemical information with measured spectra for co-analysis in order to characterize the spectra and make use of all available spectral information. This complements our use of specific peak amplitudes that are known to represent relevant chemical fragments in some cellular components, although it does not take into account unassigned peaks, peak shapes, or characteristics between peaks. MuSCA allows the co-analysis of spectral and biochemical data at each wave number and therefore emphasizes the spectral changes that may be related to their biological origin. Thus, the advantage of understanding how spectra relate to information obtained from sampling and offline analyzes is that spectral measurements can substantially reduce or completely replace these procedures.
One might also want to include a number of non-spectrometric measurements to examine all relationships between spectrometric and non-spectrometric and even mutually between non-spectrometric. One benefit that would derive from this approach is that unexpected and potentially useful relationships can be discovered.
Another recent article details the applications of Raman spectroscopy in the development of human cell therapies (3), providing an insight into the current state of emerging therapies. What advantage does the use of Raman spectroscopy offer for the analysis of individual cells or tissues?
Many different types of therapeutic cells are being developed in the hope of offering much improved treatments or even cures for various human diseases. Some cell therapies are already approved and many more are currently being tested in clinical trials. The specific role and value of Raman spectroscopy depend entirely on the specifics of the cells and the manufacturing process. Raman may be useful during therapeutic cell development by providing information about the metabolic or developmental status of cells under particular experimental conditions, or to evaluate the quality or functionality of cells as a final product. For example, we are collaborating with our co-author, Professor Timothy Kieffer, who is working on an improved protocol to direct the differentiation of human pluripotent stem cells into insulin-producing pancreatic cells to treat type I diabetes. The current protocol includes seven stages of culture in defined media where the cells are gradually induced to develop into the desired pancreatic cells. Each stage of the process involves costly culture medium and, for patient safety, it is extremely important to be able to monitor cells and determine early if there are signs of abnormal (off-target) cell development. It is also important to determine what type of off-target cells may be emerging. One aspect of our work on this collaborative project is to develop Raman spectroscopic methods to address these two goals.
After development, Raman can also be useful as a process analytical technology (PAT) for quality assurance / quality control (QA / QC) that controls the manufacturing process. Raman spectroscopy is already used as a PAT for the manufacture of other biological products (such as therapeutic antibodies) and small molecule therapeutics. The complexity of therapeutics …