https://pubs.acs.org/doi/10.1021/acssensors.4c01732
The production of high-quality biotherapeutics increasingly relies on sensors to monitor the process. One effective approach is Raman spectroscopy. Recently, Dr. Ishan Barman, a mechanical engineering professor at Johns Hopkins University in Baltimore, and his colleagues highlighted that Raman spectroscopy “has revolutionized the field of chemical biology by providing detailed chemical and compositional information with minimal sample preparation.”
These researchers enhanced conventional Raman spectroscopy in a way that could make this sensing technology even more valuable for bioprocessing applications.
Raman spectroscopy
Raman spectroscopy works by using laser photons that interact with a sample’s molecules, causing them to vibrate. These molecular vibrations alter the wavelength of the scattered light, which can then be analyzed to extract “biochemical composition information,” as described by Barman and his team. However, obtaining this data is a slow process, making Raman spectroscopy a low-throughput technique—an undesirable trait for sensors used in bioprocessing.
Compressive sensing
Despite this limitation, the process can be accelerated using compressive sensing. As Barman’s team explained, this method “enables the collection of sparse data in real-time, which can then be computationally reconstructed, often utilizing remote servers or cloud computing.” As a result, compressive sensing can enhance the speed and affordability of Raman spectroscopy while also improving portable instruments that could be deployed directly on a bioprocessing line.
To demonstrate this approach, Barman’s team applied compressive sensing to Raman spectroscopy to estimate the concentration of bovine serum albumin in water. Even when 90% of the data was missing, the technique still produced concentration predictions with “the highest coefficient of variation of around 2.7%,” according to the researchers.
Advantage in speed
Based on these findings, Barman’s team concluded: “For concentration prediction of known components, such as in the case of bioprocess monitoring, there is a clear opportunity to produce a huge advantage in speed with only a modest compromise in regression metrics.” By refining Raman-based sensing, bioprocessors could benefit from faster data collection and reduced costs—key improvements for optimizing biotherapeutic production.
Bron: www.genengnews.com