The very first SNP in a table on that study (GPBAR1
rs199986029) has a global allele frequency of 0.000004081 in gnomAD. Literally only 1 out of 245,032 didn't match reference. 1 heterozygote worldwide in this database.
https://gnomad.broadinstitute.org/variant/2-219128356-G-C
If you combine the other frequency databases, up to 4 in the world have it. But the same people can be in different databases. So it's probably 2 people total.
This was literally one of the first things I examined in the study. I kept looking to see if I typed the rsID wrong.
I agree with you completely and had noticed this. With only 3 people worldwide having this variant, it clearly is not the case that 70 percent of those with CFS have it.... My guess is that this is a bad call on the part of 23andme. This happens occasionally with 23and me data. another example is another actually from their list of differences CYP2D6. For a long time, 23andme mistakenly showed many people had a variant here that in fact they do not. NOt sure if it has been corrected yet, but likely many of those participants data got included in the study.
The question becomes why on earth wasn't this error with GPBAR1 caught? Instead they go on to intepret it as if it were a real effect! I mean, if it was a real effect, then you've got a smoking gun here and THE cause for CFS! NOte that all their subjects come from 23andme, but their controls come from already published normal reference stas like 1000 geneome project. Since 23andme makes plenty of mistakes, one really needs the control data to come from 23andme as well. That should not be hard to do.
And even if they were going to use the published normative data, still, someone should catch errors like this.
i just saw another poster said they had undergrads analysizng the data. that makes sense and it must be something like that. I'm often generally suspcious of papers with many authors, even when there areen't undergrads involved.
alas, means pretty much one needs to discount everything that is said. the oly way to rescue is if someone were to take those 50 ssnps and check frequenies agains non CFS folk - but only those tested on 23andme (and with the same chips).
one can howerver take note of the software they used and generally repeat properly.
I recently started reading klimas' papers and finding them very uneven. Iread one that was terrfic on antibodies to a protein compenent of herpes viruses in those with cfs and gulf war syndrom, so eagerly started wroking my way through papers- only to find they vary so much in quality i might as well be choosing articles at random, so wil stop my efforts of reading thru papers with her name on the authrorship. myabe will look for any with her name as first or last author.
sorry for this long post. Just wasted my time looking at their genetic data. i wil spend just a tiny bit more and see if there's anything i can mine from their supplementary table, if they have some raw datra.
oh, one more thing- the CAMM score is just silly. they use that as the primary metric of whether something is likely to be deletrous- anything above 20 - nonsense, that's pretty close to the mean CAMM score of loads of on-offs.
