Selecting individual virus particles in micrographs, or boxing, is done frequently by hand, but the automation of this process would allow for greater efficiency during the process of 3-D reconstruction. Two methods of semi-automated boxing, autoboxing and autoboxing from references, were tested, both of which are two automatic boxing settings found as a part of the program Electron Micrograph Analysis (EMAN), version 1.9. Autoboxing from references appears to pick up slightly less undesirable particles, each of which were reviewed manually after automatic selection. After boxing, selected virus particles can then be used to form a 3-D reconstruction of the virus, whose accuracy is dependent on the quality of particles selected, and help with both virus analysis and virus-like particle (VLP) creation. Particles selected from autoboxing from references, however, appear to form less accurate reconstructions of viruses after 3-D reconstruction and 10 cycles of refinements.
Examining Automated Methods of Boxing in EMAN for 3-D Reconstruction
Dr. R. Holland Cheng