If we've learned anything, we should know that people will be terrible enough to want to "cure" homosexuality or detect it before birth, like trisomy 21, if it ever will become possible - and research like this will lead to that.
I get the concern for potential risk around selective abortion, IVF, etc; I was around for the "gay gene" debate in the 90s. This isn't that. This disproves that. And what they're reporting could not be used in that way.
On that note, if you haven't read Randy Shilts' And the Band Played On, I strongly recommend it. The events it describes are, I think, the single strongest influence on everything about US (at least) gay culture that's thus far followed.
In that lens, imagine a scenario where a woman is happy to be pregnant and excited about becoming a parent. Then she learns the child will be male, and she decides to abort it. There's a wide gulf between between the harm of "didn't want to be a parent" and the disappointment of "didn't want a son". It may help to imagine this with even more trivial features: "oh the baby is going to have blue eyes? never mind then...".
For something like trisomy-21, I find myself of mixed mind. On one hand, there are many things that can happen to a child through their life that could end up posing a similar impediment to the child and burden on their parent. On the other hand, since we're able to detect it so early, perhaps it is the best way to minimize harm.