Well the purpose of my testing was to compare apples to apples as someone with little or no experience using this kind of workflow as to expectations and predictability. AKA someone who does not know what they are doing or how to go about fixing things. As I mentioned G3 had issues up until 4.15 at which point I have to believe that Epic fixed the usability problems engine side and along with material behaviour at which point Daz Studio along with Genesis 3 became a usable product as far as a pipeline tool goes. I have no problems recommending to someone starting out as in a 1-1 import as the result can be improved as to what is required by UE4 in producing a AAA result via effort rather than having to deal with problems that can not be resolved, or should be, via a graphics hack.
So
The setup for the test was to take each of the framework assets, pose them as you suggested, and export them with out further processing and as compared to each other the results were totally different from one another to the point that G8 was unusable imported into 4.25.1. As an extension to the testing I also tested both using injectors, which is a key feature of using Genesis in the first place, and the problem become worst at which point there was no need for further testing as to usability except maybe the challenge of figuring out where the problem is occurring.
I forgot to mentioned that when I applied a full body morph holes opened up under the armpits.
My conclusion is the same as yours to put G8 aside until a “key” feature is fixed on either end as the end user should not be expected to hack around in an effort to force an asset to work as to expectations. If the case this is the point where it’s better to start from scratch rather than letting a 3rd party attempt to solve the problem for you 
At the very least it’s worth pointing out the problem to Epic as they did respond to G3 problems and continue to improve the fidelity import and maybe one day fix the eyes issue (nudge nudge)