Gravity, as I see it, is the inversion and/or scattering of particle uncertainty when highly interacting particles clump together. Physicists would say that any object still maintains the possibility that it could fly part at the electron shell level (theoretically) at any moment. Think if it as a massive object's footprint spreading forward into time. Less massive objects (like you and I) near the surface of the larger simply become statistics in the largest worst-scenario desegregation, and the probabilities of where we end up are much less complex than the massive core of a planet.
The smaller, brighter masses of the continuing galactic parts of collision mentioned in the article were likely sling-shotted through because of the massive gravitational pull of the early dark matter interference, which likely significantly subsided in strength as the dark matter uncertainty was reigned in by the pull of the gravity of the normal matter. Why? Since normal matter can interact with so many more particles/forces, It's level of uncertainty is reduced by the number of nearby particles which could interfere with it. Therefore, it's gravitational footprint would remain relatively static and condensed. This would cause the dark matter (by this time lumped as observed above) to be pulled like a jetstream as the densely certain mass imposed just a bit more certainty on the dark matter around it, reducing the net gravitational pull.
TL;DR; I propose that Dark Matter's low rate of interaction gives rise to gravitational interference. As an echo of quantum uncertainty, it gives rise to large fluctuations in gravitational fields acting on normal matter. Normal matter counteracts this gravitational interference through via higher levels of certainty imposing on dark matter. Uncertainty seems to be a property of particles most purely interacting with forces.