A novel fluorescence-microscopy-based image analysis method for classification of singlet and doublet latex particles is demonstrated and applied to a particle-based immunoagglutination assay for quantification of biomolecules in microliter-volume bulk samples. The image analysis method, verified by flow cytometric agglutination analysis, is based on a pattern recognition algorithm employing Gaussian-base-function fitting which allows robust identification and counting of singlets, doublets, and higher agglomerates of fluorescent microparticles. The immunoagglutination assay is experimentally modeled by a biotin-streptavidin interaction, with the goal of both theoretically and experimentally investigating the performance of a general immunoagglutination-based assay. For this purpose a theoretical model of the initial agglutination kinetics, based on particle diffusion combined with a steric factor determined by the level of specific and nonspecific agglutination, was developed. The theoretical model combined with the experimental data can be used to optimize an agglutination-based assay with regard to sensitivity and dynamic range and to estimate the affinity, receptor surface density, molecular and binding site sizes, and level of nonspecific binding that is present in the assay. The experimental results are in good agreement with the theoretical model, indicating the usefulness of the model for immunoagglutination assay optimization. (C) 2004 Elsevier Inc. All rights reserved.
