For clarity PloS One is legit, but it's something to be a bit wary of using as a layperson. Kind of the point in PloS One is that it accepts a broad range of studies from... all the disciplines, so long as they 'contribute to the scientific literature'. They're not always good contributions (and they've had some significant issues at times), but if the methodology is clearly described and the analysis is coherent, they may still get published. So a study with a pretty crap data source (which I think this is) may still be interesting to someone provided those issues are clearly described in the paper (which they are).
Put another way anyone basing a decision on their pet's nutrition on a single study with a sample size of ±100 which specifically targeted vegan pet owners on facebook and uses self-reported data is a fucking idiot. But it's still quite an interesting study that may justify further research.