Whole genome sequencing uncovers a misclassified case of glycogen storage disease type 13 previously diagnosed as ME/CFS, 2021, Brown, Younger et al

[Simon M]

I was assuming that a GWAS would produce data that could be used to identify the rare variants that do cause 'genetic' diseases- diseases pretty much directly caused by a genetic variant (along with, of course, its main purpose of identifying genetic variation that might create a vulnerability to ME/CFS).

as I said in my previous post, rare variants are not usually like genetic diseases. I think there are rare variants that increase the risk by 30%, but that's a far cry from causing disease. The poster child for rare variants is the APOE epslion-4 variant, discovered in one of the first GWAS , and people who are homozygous for that have a substantially increased risk of Alzheimer's disease (not sure how much). But I don't think has been anything quite like it since.

I think it's unlikely that a GWAS will discover wholly-genetic diseases in ME/CFS.

So, are you saying that a careful search of the literature at this time is unlikely to find any wholly genetic diseases that produce symptoms that might possibly be mistaken for ME/CFS in medical systems that don't put a lot of effort into diagnosis?

Apart from the problem that there are unlikely to be any wholly-genetic diseases, ME/CFS comes with an incredibly wide range of symptoms. I imagine a lot of genetic diseases would have overlap with these, as would a lot of non-genetic diseases; I don't know how meaningful it would be.

That said searching the literature for genetic diseases that could present as ME/CFS could be a useful approach.

Wholly-genetic disease almost always affect the protein sequence and the protein itself. If they were genetic diseases that looked a lot like ME/CFS (though I think we have heard of them by now), one possibility is that people with ME/CFS don't have the mutation seen in genetic disease, but they do have other DNA variants that affect the same system to a lesser degree. And certainly the GWAS results could be probed for that.

Though it's not a simple process to find the link, because so much of the genome is undocumented and quite a lot of work is involved in going from the finding a significant DNA variant, to finding out what the hell the DNA variant does. (Not least because the DNA variant discovered in GWAS is usually acting as a marker flag, and the actual DNA difference that links disease is somewhere else relatively nearby).

Hope this helps. Disclaimer: I am fairly up on GWAS, but I'm not a professional, so bear that in mind!

 

[wigglethemouse]

Though it's not a simple process to find the link, because so much of the genome is undocumented and quite a lot of work is involved in going from the finding a significant DNA variant, to finding out what the hell the DNA variant does.

I spoke to a researcher about this who gave a real world example. In that case it took the researcher about a month from getting the blood sample of an ill patient to finding a rare variant of interest that could be the cause. But it took a team of researchers 3 years to prove it. You ideally need an animal model or a tissue model for proof but even then a method may not exist to measure the protein or enzyme accurately.

I also remember skim reading the guidelines for the UK Biobank GWAS statistical analysis several years ago. You have to discard rare variants because the analysis does not have sufficient sample size. I think it may have been something like values below 0.1% allele frequency should be discarded but it may have been higher. Obviously it depends on sample size. It is easy to make a mistake and get false results if you do not do this.