Regular matter apparently does effectively "clump" because the visible parts of galaxies apparently often do merge into one galaxy when they collide.
While it is true that few if any stars actually run into each other in such a galactic collision, the net effect of a large number of N-body star gravitational "collisions" (or interactions without touching) still allow the transfer of momentum between the 2 large masses of stars.
Viewed from a distance, collisions between 2 stars look the same in transfer of momentum as what occurs when 2 billiard balls hit each other. But, when there are lots of stars, not just a pair, the net gravity of the whole group tends to keep the stars from just scattering away from the galaxies. So, the net effect is that the 2 groups of stars tend to average their net momentums through intergalactic space and merge together.
Now, what that means for "dark matter" that we cannot directly see is going to depend on what that dark matter really is. If it is "another world" of matter with "dark atoms", "dark stars" etc. then I would expect it to have the same merging effect as the visible matter. But, if it is particles or waves of some sort that do not interact with each other the same way that regular matter interacts with itself, then the dark matter may do something entirely different.
But, if it really does just "pass through", it still should be attracted to the visible matter that stops at the merge volume. So, in that speculative "artist's illustration" where it seems to be attracted out ahead of the visible matter as the separate galaxies approach each other, I would expect it to also slow down as it moves away from the merged visible matter when it gets to the other side and is moving away.
If it does not end up staying with the visible matter, then there should be some sort of dark-matter-only clumps sailing around the cosmos from previous collisions, which should be detectable by their lensing effects on background light from visible matter. Do, we see anything like that - lensing with no visible foreground object to account for it? Considering that dark matter is supposedly the majority mass of galaxies, even including the central supermassive black hole, the lensing of a blob of dark matter with no visible matter in it and no black hole in it should still make a substantial lens.