Pantropy during the early stages of terraforming, then it just sticks.
"As of 1 December 2022, there are 5,284 confirmed exoplanets in 3,899 planetary systems, with 847 systems having more than one planet."
Presumably humans would start colonizing "first-generation" worlds that are both relatively nearby (relatively easy to travel to) and easy to transform (and not already filled with alien species). Once those are occupied (by Earthlings or alien species), humans would start colonizing worlds that are more distant or more difficult to transform or both.
As o.m. suggested,
Say you have a world that should be shirtsleeve-habitable for
unmodified humans in just, oh, three or four centuries. So the
colonists alter their offspring to meet the terraforming project
halfway. A modified colonist will be able to survive outside in a
century or so. Much better for their mental health that they could go
out, in case a seal breaks. And by the time a baseline human could
survive outside, will the colonists change back? (Various writers have
expounded on how supposedly the first generation builds things, the
second maintains them, and the third lets maintenance slide. So it is
a bad idea to rely on domed cities for centuries.)
Let us assume that most planetary-mass objects under consideration for terraforming are on a linear spectrum of high-gravity world, therefore dense atmosphere, therefore high oxygen on one end, vs. low-gravity, therefore low oxygen at the other end, with Earth at 21% oxygen somewhere in the middle. (I feel this is plausible enough and useful for plot reasons, although there are likely many other variables that can't be condensed down into a single linear spectrum).
Assuming space travel is reasonably easy
(especially if the more extremely modified colonists can't survive in conditions that are shirtsleeve-habitable for unmodified humans),
after the first "generation" of worlds are completely terraformed (or perhaps 50% terraformed),
and colonists begin to be selected to send to the second "generation" of worlds to be terraformed,
it seems likely that second-generation low-oxygen worlds will have more volunteers from low-oxygen first-generation worlds, and second-generation high-oxygen worlds will have more volunteers from high-oxygen first-generation worlds.
Even if space travel is difficult,
the genetic engineering will probably be easier / more successful starting from the descendants of already-modified colonists and making smaller tweaks, rather than starting from scratch from baseline humans every time.
If space travel is very easy,
some colonists modified to thrive in, say, 15% oxygen levels
may decide to travel for the first half of their lifespan, helping terraform even lower-oxygen planets
(where they only need the equivalent of scuba gear to work outside, rather than the bulky spacesuits required by baseline humans or colonists modified to thrive at 50% oxygen),
then later "retire" on some currently-near-15% oxygen world that is shirtsleeve-habitable to them (or will reach 15% oxygen / shirtsleeve-habitable in a few years).
After a few "terraforming generations" (likely each one several human generations), there will be a spectrum of modified humans.
Also, at any one time there will be a spectrum of worlds inhabited by human descendents:
- worlds slightly past the ends of the spectrum of "shirtsleeve-habitable" by any human descendent that are being actively terraformed to bring into the modified-colonist range;
- worlds that are technically inside that spectrum, but are still being actively terraformed because it is still "relatively easy" and would allow a greater variety of plants and animals to grow;
- worlds that for a variety of reasons are not currently actively being terraformed, but are currently inside that spectrum (possibly because of terraforming activity in the past that was halted before it became habitable by baseline humans).
(this builds on o.m.'s answer).