Procedure for Detection of Membranes in Three-Dimensional Subcellular Density Maps

Abstract

Electron tomography is the leading technique for visualizing the cell environment in molecular detail. Interpretation of the three-dimensional (3D) density maps is however hindered by different factors, such as noise and the crowding at the subcellular level. Although several approaches have been proposed to facilitate segmentation of the 3D structures, none has prevailed as a generic method and thus manual annotation is still a common choice in the field. In this work we introduce a novel procedure to detect membranes. These structures define the natural limits of compartments within biological specimens. Therefore, its detection turns out to be a step towards automated segmentation. Our method is based on local differential structure and on a Gaussian-like membrane model. We have tested our procedure on tomograms obtained under different experimental conditions.