Ab initio random model method facilitates 3D reconstruction of icosahedral particles

Abstract

Model-based, three-dimensional (3D) image reconstruction procedures require a starting model to initiate data analysis. We have designed an ab initio method, which we call the random model (RM) method, that automatically generates models to initiate structural analysis of icosahedral viruses imaged by cryo-electron microscopy. The robustness of the RM procedure was demonstrated on experimental sets of images for five representative viruses. The RM method also provides a straightforward way to generate unbiased starting models to derive independent 3D reconstructions and obtain a more reliable assessment of resolution. The fundamental scheme embodied in the RM method should be relatively easy to integrate into other icosahedral software packages.

Averaging tens to hundreds of icosahedral particle images to resolve protein secondary structure elements using a Multi-path Simulated Annealing optimization algorithm

“…For instance, the multi-refine program has been used in EMAN, where several initial models of different conformations or multiple models with varying amount of added noise would give rise to the same final map (Brink et al, 2004;Chen et al, 2006). Alternatively, Yan and coworkers (Yan et al, 2007) proposed the Random Model (RM) method and showed that random models can be effective starting models. They split whole dataset into two halves to generate two independent models (starting from two individual random initial models) which have to agree with each other.…”

Structure of the archaeal head-tailed virus HSTV-1 completes the HK97 fold story

“…The digitized micrographs were processed, and viral particles were picked as previously described (49). The initial models for the viruses at approximately 30-Å resolution were generated using a random model computation method with a subset of 150 far from focus particles (50). These models were then used as starting maps for iterative full orientation searches and origin determinations of all of the particles using AUTO3DEM (51) including periodic recentering of the particles using RobEM.…”

Structure Determination of Icosahedral Viruses Imaged by Cryo-electron Microscopy

“…Geometric models of the proper dimension can also be used, but again are prone to the same problems. An alternative approach that we typically employ is to use the 'random model computation' (RMC) 20 to construct a starting model from a relatively small number of particle images.…”

“…In our reconstruction scheme, we generally use AUTO3DEM to construct an ab initio model using the RMC. 20 The crux of the RMC approach is to construct a density map from a small number of particle images for which random orientations are assigned and whose origins are set to the center of the box. Although this initial 3D map will bear no resemblance to the actual structure being solved except for appropriately representing the size and symmetry of the virus, it often serves as an effective seed for the image reconstruction process.…”

“…FSC data from iterations 0 to 4 are so noisy that resolution estimates are considered to be unreliable, but suggest that the maps are at resolutions worse than 30 Å . 20 …”