Molecule binding parameter estimation from Surface Plasmon Resonance measurements

Surface Plasmon Resonance (SPR) is a highly sensitive technique that utilizes the properties of surface plasmons to detect the subtle changes in mass caused by adsorption or desorption of molecules. The SPR technique is appealing because the amount of adsorption or desorption is detected in real time, without the need to label the adsorbate or prepare samples using a complex procedure. Although much progress has been made in the analysis of SPR images several challenges remain. In this paper, we addressed two of these challenges, 1) automatic detection of regions of interest (ROIs) and 2) accurate estimation of the molecular association and disassociation parameters. With hundreds of ROIs on a single SPR video frame image, our procedure to automatically detect the ROIs greatly reduces the labor and time. The gray level values of the ROIs were extracted over time and used to estimate the molecule binding parameters, which are vital in biosensing applications. The parameters were estimated using Particle Swarm Optimization (PSO) and the standard Levenberg-Marquardt (LM) algorithm. A comparison of the results revealed that the PSO algorithm achieved a much lower mean squared error (MSE), and hence more accurate, than the LM algorithm for all of the active ROIs.