Quantitative analysis of 3D alignment quality: itsimpact on soft-validation, particle pruning andhomogeneity analysis

Abstract

Single Particle Analysis using cryo-electron microscopy is a structural biology technique to capture the three-dimensional conformation of biological macromolecules. The projection images used to construct the 3D density map are characterized by a very low signal-to-noise ratio to minimize radiation damage in the samples. As a consequence, the 3D alignment process is a challenging and error prone task and this job usually determines the success or failure of the macromolecule reconstruction. In this work, we present a soft-alignment validation approach, which can quantify the alignment precision and accuracy as well as the data homogeneity of the single particles when they are confronted with the resultant reconstructed 3DEM map. We have also applied this method to data homogeneity analysis and particle pruning, improving the data quality and as a consequence the final map resolution.