This isn't a fully thought-out answer, sorry if it rambles a bit (anyone who wants to edit or add, feel free), but I just wanted to briefly draw attention to how we treat questions that fit other sites.
For questions that need answers based purely in biology or physics, without any special biological consideration, we tend to close or migrate those questions out. It isn't sufficient for the original motivation to be based on biology if the question isn't. Same for questions that are purely about statistical approaches, even if those approaches are more relevant to biology, although we do have many questions that we allow that are specific to understanding evolution/natural selection/Mendelian genetics. What distinguishes those questions is that they have a biological question at heart, and statistics just helps you to understand the biological question, rather than the question being "how do I control for a confounding variable in this ANOVA?" where the question isn't actually about the biology.
For questions that could be better fits somewhere like Health.SE or CogSci.SE, we sometimes encourage the OP to delete and repost, but we usually keep them open and try to answer them, as long as they also fit the on-topic criteria here. Even if they might fit better someplace else, they can still be answered here.
For Bioinformatics, I think part of the reason there is a new sub now is that 95% (or 99.9% - calling @GerardoFurtado!) of the bioinformatics questions people actually ask are not on topic for Biology even though we have a tag here. I think future Bioinformatics questions will continue to be welcome here, even with Bioinformatics in beta and even if it graduates, much like the CogSci and Health questions are: as long as they are actually about the biology.
We can inform and encourage askers that there is another SE site, but I don't think biology-relevant bioinformatics questions should become off-topic here.