Polarized Raman spectroscopy for enhanced quantification of protein concentrations in an aqueous mixture

Raman spectroscopy (RS) for selective quantification of protein species in mixed solutions holds enormous potential for advancing protein detection technology to significantly faster, cheaper, and less technically demanding platforms. However, even with powerful computational methods such as nonlinear least squares regression, protein quantification in such complex systems suffers from relatively poor accuracy, especially in comparison with established methods. In this work, a combination of the expanded set of spectral information provided by polarized Raman spectroscopy (PRS) that is otherwise unavailable in conventional RS was, to our knowledge, explored to enhance the quantitative accuracy and robustness of protein quantification for the first time. A mixture containing two proteins, lysozyme and α-amylase, was used as a model system to demonstrate enhanced quantitative accuracy and robustness of selective protein quantification using PRS. The concentrations of lysozyme and α-amylase in mixtures were estimated using data obtained from both traditional RS and PRS. A new method was developed to select highly sensitive peaks for accurate concentration estimation to take advantage of additional spectra offered by PRS. The root-mean squared errors (RMSE) of estimation using traditional RS and PRS were compared. A drastic improvement in RMSE was observed from traditional RS to PRS, where the RMSEs of α-amylase and lysozyme concentrations decreased by 11 and 7 times, respectively. Therefore, this technique is a successful demonstration in achieving greater accuracy and reproducibility in the estimation of protein concentration in a mixture, and it could play a significant role in future multiplexed protein quantification platforms.