Genetic engineering is a nearer and more realistic goal than Mars colonization, so I think it'd be wise to assume we achieve the former before the latter and plan accordingly.
"Genetic engineering" is not a single, monolithic technology; we don't go from putting GFP in rats and running somatic cell nuclear transfer on sheep to suddenly being able to make whatever we want, regardless of the pop-sci hype about CRISPR that's got all the laypeople excited. Nor is it reducible to a series of modular, independent steps. You can build, say, a lander to get a payload of mass X onto the Martian surface from orbit and then have some other people build a cruise stage to get that lander across interplanetary space while treating the lander as a black box of a certain mass and power requirement and so forth. Systems biology doesn't really work that way; everything you change interacts with everything else you change, and cells are surprisingly intolerant to metabolic stresses of the type induced by expressing exogenous proteins. Add in too many (read: about a dozen, depending) and they get sick and die even as immortalized cell cultures, let alone as, you know, primary cells in an actual organism. So the usual naive solution of giving people some capability they're vaguely aware some other animal has, be it a bear's ability to hibernate or a mole rat's resistance to cancer or anything at all about tardigrades, would involve reengineering the entire proteome to compensate for the difference in amino acid usage at minimum and we don't know how to find the mechanisms controlling some of that, let alone tune them, and that's only if the differences are purely chemical. Anything anatomical means figuring out not only what we want but also how to get cells to communicate how to build it. The human genome doesn't contain, say, a "gene for bones" that we can just edit to make them universally lighter and longer and thinner. It has a whole welter of genes for guiding that development process in ways that are controlled quite literally mechanically as well as chemically in countless molecular contexts. Again, we don't know where a lot of those controls are, let alone how to tune them.
The single biggest problem, though, is the sheer amount of time this would all take. Humans take forever to grow, and even if you've figured out two changes you want to make it's going to take at least another generation to see if they can both be made in the same organism -- and those are changes like tuning the expression of individual proteins, not changes like "low gravity tolerance" generally.
Putting humans on Mars is a laughably difficult problem, but it's made up of smaller problems that can, within reason, be dealt with individually. We can enumerate everything we can't do and where the shortfall lies. (Mostly it's a bunch of little things with environmental control and a literally massive problem with our limits in getting heavy equipment onto the surface exacerbating a number of logistical challenges.) Genetic engineering on the scale you propose, though, involves making thousands of regulatory elements we don't yet know exist work in concert with physiological changes both subtle and gross in ways we cannot yet accurately model, and they must all be made to work together. That's a much more difficult problem even taking into account that it'd take a literal lifetime to know whether or not we succeeded.