Assuming the strata theme is fully deterministic (nothing random, you can follow a set of rules to reach the result), then my observations mean either the order of deposition matters, or some mysterious factor is at play (like absolute number of minerals, time of deposition, etc.). I am more inclined to believe in the former, that the order in which the final 5-4-3-2-1 permutation is achieved affects the resulting strata, when the first, second or more combinations gets anti'd. At first blush this theory might account for why different strata themes could arise from the same permutation, though I admit I haven't sat down to think it out yet. If this is the case, testing will be tricky, as I do not have large reserves; I'll have to watch a new gate fill up starting from zero.

Of course, if the formula for anti'd strata is not deterministic, then there's no point in trying to figure this out. (i.e. if the primary combo is anti'd a random theme is picked)