I think your argument starting with
> For instance, you cannot use genetic studies to define a race
is weak. For millennia we had phenomena that science couldn't describe. For many different reasons. One can now say that due to this the phenomena simply don't exists, since they didn't had a sound scientific understanding base. But that was obviously wrong.
To give you one of thousands example, think of how long it took science e.g. to understand lightnings. You could analyze the lightnings by form, and you ended up with thousands of forms. And then you could just say "Oh, this is so unscientific (because by some arbitrary limit reached), in reality lightning doesn't exist".
If some phenomena 1) isn't yet 100% scientifically described, then this is perhaps just a sign that science isn't far enough, hasn't used the proper methodology?
This is BTW a reason why I like the german word for science better than the english one. In german, it's "Wissenschaft" -> "knowledge creation". It's not "Wissenhab" -> "knowledge having". The word itself, unlike the english one, transports the idea that our knowledge is still limited. And that science is a process, not some achieved state.
1) "race" isn't just some old-white-man western concept. If you go to China, then you see that they also has such a concept --- and a quite strong one about "Han chinese". Or some african tribes, they even thought wars (Tuts & Hutu in Ruanda). Maybe race isn't a genetic concept, but a sociology one, e.g. because humans for whatever reason want to have an "us and they" thingy. In such a case searching for race genetics is simply the wrong methodology. Rendering the approach invalid, but nothing else.