I majored in Statistics in college (with a minor in English Lit), but my stats sophistication is a bit rusty now. But it's not so rusty that I do not raise my brow at any latest stats reported in medicine, or especially in Psychiatry - and especially genetic studies.
As Gene Expressions points out, it's partly because a p-value of 0.05, commonly used in such studies, is unrealistic for these things. It's straight out of How to Lie with Statistics, which is essential reading for all high school students.
As the man says, if there is a genetic serotonin link with trauma and depression, it has yet to be proven.
In his second post on the topic, Why are most genetic associations found through candidate gene studies wrong? he makes the key point:
For nearly all diseases, reproducible associations have small effect size and are only detectable when one has sample sizes in the thousands or tens of thousands (for many psychiatric phenotypes, even studies with these sample sizes don't seem to find much). The vast majority of candidate gene association studies had sample sizes in the low hundreds, and thus had essentially zero power to detect the true associations. By the argument above, in this situation the probability that a "significant" association is real approaches zero. The problem with candidate gene association studies is not that they were only targeting candidate genes, per se, but rather that they tended to have small sample sizes and were woefully underpowered to detect true associations.
While I find the field of behavioral genetics to be as fascinating as anything else in this world, I always read the latest gene-behavior studies with the highest skepticism. (Do I think real Bipolar Disorder has some provable genetic underpinning? Yes, I do, even though I do not think it has been adequately proven yet. But not much else genetic in Psychiatry has been adequately proven in my view. Schizophrenia maybe, IQ almost certainly, but possibly not homosexuality, or depression, or alcoholism.
The trick to getting papers published is to run your numbers so they show something. It's not rocket science if you know how to do it: just look at the climate studies. (Even Einstein fudged his math. He happened to turn out to be right, though, as far as we know today.) Science is about hypotheses, not Truth.
