Determining Orientation in Cryoem Single Particle Analysis

Abstract

In single-particle electron microscopy, the electron dose is limited to avoid damaging the specimen. This results in images with low signal-to-noise ratios (SNR). Class averaging techniques are used to enhance the low-SNR electron micrograph images. A class is defined as a collection of projection images taken along nominally identical projection directions. The images in each class are aligned and averaged in order to cancel or reduce the background noise. The class-averaged images with high-SNR can be used for more accurate threedimensional reconstruction in single-particle electron microscopy. However, errors in the alignment process are inevitable due to noise in electron micrographs. This error results in blurry averaged images. Using the mean and variance of the background noise that is assumed to be Gaussian, we derive equations for the mean and variance of translational and rotational misalignments in the class averaging process. Furthermore, the blurring function representing the distribution of the misalignments is estimated using a Gaussian with the computed mean and variance of the misalignments. The blurring process in class averaging is formulated as convolution of an underlying clear image with the blurring function. We propose a deconvolution method to estimate the underlying image using the Fourier analysis in the appropriate domain. This deconvolution method is applied to artificial and experimental electron micrographs. The deburred class averages are assessed quantitatively and qualitatively.