If you were to calculate the average B factor of the subset of atoms in the protein that are exposed on the surface you would find that this is higher as well. Most solvent atoms are on the surface of the molecule and everything has higher mobility there.
I would be suspicious of a model where the average B factor of the solvent wasn't higher than the protein. Dale Tronrud On 08/11/11 07:10, David Schuller wrote: > At the request of reviewers, I worked up the average B factor for some > structures using CCP4 programme BAVERAGE. Here's one example: > > Protein 33.1 > Solvent 53.9 > Ligand 32.9 > > I think the reviewer's question was directed at the quality of the model > for the ligand, but I notice that the average B for solvent (mostly > water, but also glycerol and sodium) is quite a bit higher. This was > refined in REFMAC5 with anisotropic thermal factors. Does REFMAC5 place > restraints on water that inflate its B values? >
