A Monte Carlo framework for missing wedge restoration and noise removal in cryo-electron tomography

Emmanuel Moebel,Charles Kervrann

doi:10.1016/j.yjsbx.2019.100013

Abstract

We propose a statistical method to address an important issue in cryo-electron tomography image analysis: reduction of a high amount of noise and artifacts due to the presence of a missing wedge (MW) in the spectral domain. The method takes as an input a 3D tomogram derived from limited-angle tomography, and gives as an output a 3D denoised and artifact compensated volume. The artifact compensation is achieved by filling up the MW with meaningful information. To address this inverse problem, we compute a Minimum Mean Square Error (MMSE) estimator of the uncorrupted image. The underlying high-dimensional integral is computed by applying a dedicated Markov Chain Monte-Carlo (MCMC) sampling procedure based on the Metropolis-Hasting (MH) algorithm. The proposed MWR (Missing Wedge Restoration) algorithm can be used to enhance visualization or as a pre-processing step for image analysis, including segmentation and classification of macromolecules. Results are presented for both synthetic data and real 3D cryo-electron images.

Highlights

Cryo-electron tomography is generally used to explore the structure of an entire cell and constitutes a rapidly growing field in biology
We focus on the Cryo-electron tomography (cryo-ET) restoration problem but the proposed algorithm could be potentially used to address a large range of applications including medical and seismic imaging, and other inverse scattering problems
We demonstrate the potential of the method on real cryo-ET data

Summary

Introduction

Cryo-electron tomography (cryo-ET) is generally used to explore the structure of an entire cell and constitutes a rapidly growing field in biology. The sharp transition between these two regions is responsible for a Gibbs-like phenomenon: ray- and side-artifacts emanate from high contrast objects (see Fig. 1), which can hide important structural features in the image. Another type of artifact arises from the incomplete angular sampling: objects appear elongated in the direction of the blind spot (see Fig. 1), in other words the data has an anisotropic resolution (e.g. linear features perpendicular to the tilt axis disappear). The quality of tomograms can be improved if sophisticated algorithms such as MBIR (Yan et al, 2019) are applied instead of conventional methods (e.g. WPB (Radermacher, 1992), SIRT (Gilbert, 1972))

Objectives

Results

Conclusion