The resulting map for the outer RNA shell in the icosahed-rally-averaged map was aligned with that for the asymmetric RNA organization in the tomogram by reference to the X-ray protein structure used to create each difference map, via UCSF

Difference map between tomogram and X-ray protein structure

A difference map between the asymmetric EM reconstruction [6] and the X-ray structure of the protein capsid (PDBID 2MSZ) was determined as follows: the protein structure was filtered to 39A resolution to match the EM data; then the pixel size and orientation of the two maps were made equivalent by trilinear interpolation of the reduced-resolution X-ray structure with Chimera [53].

The X-ray structure provided the first definition of RNA PSs [21].

In addition to the X-ray density the modelling used predictions of the most likely secondary structure elements within the genome to identify the sequences forming the double-stranded segments [44].

In particular, many RNA viruses show order in the organizations of their genomes in icosahedrally-averaged cryo-EM and X-ray structures [46], for example Bean Pod Mottle Virus [47], STMV [30] and Pariacoto virus [48].

Symmetry averaging techniques have been used to determine viral capsid structures at atomic resolution by X-ray crystallography, and by reconstruction of such structures at medium resolution by cryo-electron microscopy (cryo-EM).

This difference arises because of technical aspects of the ways the EM and X-ray data are collected.

The X-ray structures of ppGalNAcT isoforms 2 (PDB ID: 2FFU) and 10 (PDB ID: 2D7I) were superimposed using Accelrys Discovery Studio Visualizer 3.1 to minimize RMS distance between corresponding C-alpha atoms of the catalytic domain (81 Fig).

The initial model was prepared from the X-ray structures of human isoform 2 [21] (PDB: 2FFU) and isoform 10 [22] (PDB: 2D7I), where the former includes a short acceptor peptide EA2 and the UDP part of the donor molecule, and the latter includes a hydrolyzed UDP-Gal-NAc.

Residue Included part Function agreement with available experimental evidence [12] , including the recent X-ray structures based on modified substrates.

MD simulations allow us to explore molecular motions and the dynamic interactions between lipids and proteins, providing a complementary approach to the temporal and spatially averaged static structures determined by X-ray crystallography.

So, how close is the structure of a membrane protein in a lipid bilayer determined by MD simulations to the original structure of the protein determined by X-ray crystallography in the presence of detergents?

Tightly-bound lipids have been identified in a number of X-ray crystal structures of membrane proteins showing that there are specific binding sites for lipids on the surface of some membrane proteins, which may assist in their folding or functioning [7].