> However, because our world is cosmological and the big bang does actually pick out a preferred reference frame, Lorentz Invarience has to be approximate in some way.
A particular solution to a set of equations does not usually have all the symmetry of the equations. In fact, I would say that, any solution to GR equations with energy and/or matter (assuming homogeneity on large enough scale) you will find something that looks like a preferred frame of reference.