All good points. A single, non-replicated, GWAS finding is not sufficient - I agree. This is why my personal point of view is that a well-powered GWAS (with subsequent replications) is the first thing that is needed. Only then will causal hypotheses be generated, allowing findings to be "battle hardened" afterwards. The time-scale is long: probably 3-4 years. The cost is substantial: probably £3m. But in my view (and those of others) this is the quickest approach for generating causal hypotheses that can then be followed up. With respect to this single observation from the Neale GWAS, I am more convinced than you are because the same DNA variants that predict ME/CFS risk are also associated with altered level of the ornithine transporter gene. On this, I am really pleased that two researchers - one already working in ME/CFS, one not - have been in touch. One indicates that there is much unpublished data that *might* be consistent with the ornithine transporter hypothesis. The other works on mitochondria and we will discuss soon. I wish that all of this could have been done previously, by last Christmas, and the ones before that. I do not know how to go faster.