I'm *not* a big, big fan of PLoS One's criteria for publications, but I do like their idea of publishing articles across different research areas, such as biochemistry, bioinformatics, genomics, physics, medicine, psychology, etc. The idea is that this allows researchers to not become so pigeon holed and start to see connections between their field of research and others.
It would be nice to see some other journals take this approach, although maybe not to the extreme of PLoS One. For example, for my MSc thesis, I performed a bioinformatics analysis of expression data for colorectal cancer. Instead of publishing it in a bioinformatics type journal, I got the results published in a cancer biology journal (No, that was not a shameless plug!). I think doing so allows the readers of the journal to think outside the box and see what are the other ways of tackling the same problem.