Initially non-flat space-time makes dark matter obsolete$\dots$

After the Big Bang, dark matter clusters and build the seeds for visible matter, that then forms stars, planet, you, me, everybody...

But do we really need dark matter to build cluster first. What if spacetime isn't flat per se on cosmlogical sclaes but rather crumbled or heterogenous, like a 4-dimensional version of folded cloth. We don't need dark matter, if we accept that space time didn't came along at the big bang in a clean way, as mathematical text books on general relavity suggest.

Is it possible to support this idea with astronomical data of galaxy mergers?

• What does "in a clean way" mean specifically? In the usual cosmology, the initial conditions could be chaotic and heterogeneous--but would be smoothed out due to inflation (which is necessary because large-scale homogeneity and isotropy are observed facts). And of course on the galactic scales on which dark matter is most important, the spacetime curvature is already quite heterogeneous. So I'm rather unclear about what you're trying to do... – Stan Liou Mar 24 '16 at 22:22
• Bullet Cluster en.wikipedia.org/wiki/Bullet_Cluster seems to indicate that you can segregate dark matter from regular matter. Not clear how that'd work if dark matter were really wrinkly spacetime. – Wayfaring Stranger Mar 25 '16 at 2:36
• You may be able to explain away dark energy with inhomogeneities of space-time (which we know is reality, since matter is inhomogeneous, but this is usually ignored in cosmological models). – Walter Mar 25 '16 at 18:07