26 Feb A Breakthrough in the Genetics of Anorexia Nervosa?
Julie O’Toole, of the Kartini Clinic for Children and Families, recently posted a blog with the title I have borrowed here, referring a new study of Anorexia Nervosa from the Psychiatric Genetics Consortium, led by Cynthia Bulik. Unsurprisingly I don’t think the study qualifies as a “breakthrough,” but I want to be clear about what I don’t mean here. I don’t have any particular commitment one way or the other as to psychological vs. biological explanations of anorexia, nor do I have any technical criticisms of the GWAS the group reports. In fact, like most reports out of the PGC it is beautifully done. I have no reason to doubt any of the findings they report. The paper itself is quite sensible about what it means.
Yet the paper doesn’t mean any of the things that O’Toole thinks it does. Apparently she has been asserting for some time that anorexia was heritable and a “biological disorder.” She opposes a list of psychoanalytic-sounding explanations of why some people are anorexic. She thinks she has been vindicated by the new GWAS, which identifies a single SNP at genomewide significance, computes a SNP heritability of .2 and documents a set of genetic correlations between anorexia, other psychiatric conditions, and normal eating behavior.
The heritability of anorexia has been known to be considerably higher than .2 for at least twenty years. So if that is what we are talking about– that somehow the heritability of anorexia establishes as a real biological thing, then there is nothing new here. It’s nice that GWAS and GCTA confirms the findings of the old twin studies, but they don’t establish the biological status of anorexia any more than the twin studies did. Everything is heritable, blah blah blah. The psychoanalytic (or other psychological) explanations of anorexia may or may not have any validity, but the heritability of anorexia and the occasional significant SNP association don’t have anything to do with it. Freud and GWAS could both be right.
And in fact, the GWAS reinforces how very difficult it is going to be to ever tell a biological story about something like anorexia. The paper doesn’t report a polygenic score for the prediction of anorexia, but the total effect of all the SNPs identified isn’t going to be big. In fact it is sure to be small. The pathway analyses and all the genetic correlations are interesting enough, but we know by now that there is no biological revelation just around the corner. Some people think this is the first step in the direction of biological mechanism. I don’t. When someone comes up with a surprising biological explanation of an apparently behavioral syndrome, I will be contradicted. It hasn’t come close to happening yet, and the the inexorable progress of science does not guarantee it will happen.
So that’s it. We are obviously going to proceed with the GWAS of everything, and so be it. But in the meantime it would be good to avoid making all the exact same over-enthusiastic mistakes we made twenty years ago about twin studies.