Let me just start with saying that I completely disagree, to get that out of the way.
As I mentioned in the chat I find it very odd that one's feeling can get in the way of a scientific, unbiased question. In my experience, it feels like we cannot talk about facts anymore, just because someone is sad or mad about the topic. Say someone is offended by all the "Vaccines do not cause autism" research. Must we stop discussing it then? Must we accept one's views based on one's feelings? Well the answer is simple: Never!
This is the part where I will bring some nuance in my answer. As mentioned above, scientific and unbiased questions should never be deleted because someone's feelings got hurt.
In this specific situation, the question was not unbiased. It is rather clear that the OP hoped to confirm that homosexuality is indeed a disorder. This became especially clear when the answer was accepted that gave one minor hint of a disorder (the conflict between wanting to be homosexual or not vs. actually being homosexual or not). However, despite this bias, an overwhelming amount of arguments and proof was provided against. In other words, the truth came forth (which even corresponds with societal views). Why is that an issue?
Course of action
What should we do to prevent these kinds of question to explode like this? Simple: prevent them to be so biased-laden. We have closed quite some biased questions (Reference needed) and can (and should) do so still. This allows us to make the question less bias-laden, so that we can have an open, normal, objective discussion about it. Sensitive subjects are not unspeakable subject, and if we start censoring open discussions, we'll be living in 1984.
What should we do with this specific question about homosexuality? We need to improve the question: remove bias, possibly add credible hypotheses, etc. The question is an important for science (and for CogSci), and removing it cannot be justified to the answerers who put time and effort to write up such amazingly nice answers.
This question, and its result, is exactly why you guys started CogSci: multiple answers with different points of view and different sources, either competing or cooperating.