PERCEPTUAL GROUPING OF MEMBRANE SIGNALS IN CELL-BASED ASSAYS

Membrane proteins organize themselves in a linear fashion where adjacent cells are attached together along the basal-lateral region. Their intensity distributions are often heterogeneous and may lack specificity. Grouping of these linear structures can aid in segmentation and quantitative representation of protein localization. However, quantitative analysis of these signals is often hindered by noise, variation in scale, and perceptual features. This paper introduces an iterative voting method for inferring the membrane signal as it relates to continuity. A unique aspect of this technique is in the topography of the voting kernel, which is refined and reoriented iteratively. The technique can cluster and group membrane signals along the tangential direction. It has an excellent noise immunity and is tolerant to perturbations in scale. Application of this technique to quantitative analysis of cell-cell adhesion mediated by integral cell membrane proteins is demonstrated