The correspondence cannot be expected to be 100%, as the protein flexibility is different in the crystal from the case in solution (as in the simulation). Additionally, not only flexibility is encoded in the crystallographic B-factor, but also static disorder, experimental error and model inaccuracy.