Another Unreplicated, hyped study of Anorexia Nervosa

This post takes a quick look at this GWAS study for Anorexia Nervosa. Let me start this off by saying that Anorexia Nervosa is quite clearly a disorder created by our society and compounded by particular childhood trauma issues that I won't get into here. Moreover, looking at this disorder as some kind of metabolic disease and trying to pass it off as a genetically based disorder is very unfortunate and I think shows some real lack of understanding of clinical disorders, and bizarrely reduces it to chemical mechanisms.  In the long list of authors for this paper, I would hope at least a few have some actual clinical experience with this disorder, but you wouldn't get a sense of it from the paper.
Now that I got that out of the way, let me briefly discuss the paper itself, which follows the usual formula for GWAS these days:
Take all the data from old studies, add on some new data and do a GWAS as a "meta-analysis," without ever looking at the new data independently. The advantage of this method, is that there is no way to independently evaluate any significant loci found between studies ( which doesn't even matter in this instance).  This study added about 4,000 more cases to this study, which had 13,000 cases and found one "significant" locus. This was touted heavily at the time, suggesting that this locus is also related to diabetes.
So let's consider what we might expect to find when some new data is added to the previous study and a new GWAS is performed: We would expect that it would easily "replicate", if it was a valid finding, would we not? You are given an N=13,000 head start for a significant locus, you already correlated it to diabetes, then you get another 4,000 or so cases added on, it should be a sure thing... Well, no, this locus didn't reach significance. Instead they found 8 new loci. Why would anyone expect that these 8 loci are anything more than false positives before any attempt to replicate them, when the only previous finding couldn't even hold significance with a huge handicap?
The new data is never assessed (at least not included in the study) independently of the previous data. Instead they go for "enrichment" of various CNS regions, which they link to after the fact. Again, if you didn't have this in your hypothesis, it is not a "finding". It is something that you might want to look at up front in your next study. One would think that finding enrichment in the cerebellum, for example, which would not have any logical link to Anorexia Nervosa, would be a tip off that you are playing with junk data.
So to review:
No replicated significant loci as yet, and the only previous significant locus did not replicate despite that data being used in this study.
"Enrichment" findings all come after the fact, have no obvious causal mechanism to the disorder, and have yet to be replicated (and probably never will).