The genetics of ME: A commentary on Hajdarevic et al., 2022, Ponting & McGrath (w. blog summary)

This is a commentary on the recent GWAS paper (thread) by Dr Rias Hadjerivic and colleagues (including FLuge & Mella).

As it is a commentary, there is no abstract, but you can read the full text free until Aug 5 (preprint here). I have also written a 600-word blog to summarise and explain our piece:

Norwegians publish the biggest ME DNA study yet and show we need even bigger studies

A Norwegian team has published the largest analysis yet to look for DNA differences that could pinpoint what goes wrong in ME (also known as chronic fatigue syndrome, CFS). Such differences would be a first step towards finding effective treatments.

Unfortunately, the new study doesn’t find any DNA differences that reach the accepted standard for statistical significance. Even so, as Professor Chris Ponting and I comment in a companion piece, their paper helps to move forward the field of ME genetics.

Ruling out dead ends
Study leader Dr Riad Hajdarevic and colleagues found no support for DNA differences that others had linked to ME. We feel it is critical they reported these negative results to help ME research refocus on the most promising leads.

The authors conclude that all previous genetic studies have not been big enough to find true associations between DNA differences and ME.

No new findings reach statistical significance.
The researchers ran a “discovery analysis” on 427 Norwegian patients and looked to replicate the findings from this in 2,105 UK Biobank participants. They did not find any new links of DNA differences to ME that reach the accepted threshold for significance (p < 5 x 10-8).

We agree with the authors that, to pinpoint and tease apart ME’s various causes, researchers will need to create much larger studies, such as DecodeME (which aims for 25,000 ME participants). Scientists will also need to pool the data from many studies to analyse even larger numbers of patients.

...

Studies need large numbers of participants to reliably detect individual DNA differences, which often have only a very small effect. (Note: the first ‘big DNA’ studies, which were smaller and used less-strict levels of statistical significance, usually produced unreliable results.)

The latest view is that, in most cases, studies need at least 10,000 participants to produce reliable results.

The critical role of those with ME
Finally, Chris and I say that people who live with ME should be at the heart of ME studies because their experience will always improve scientific quality and delivery. People with ME can also consent for their data to be shared ethically to allow researchers to analyse data from even larger numbers of participants. This is the fastest route to robust genetic results.

Full blog

 

Thanks Trish. I'm not sure what happened there but I have now updated the links in my post.

 

@Jonathan Edwards

You mentioned something about solid phase information and how this could explain why the mechanism in ME has been so hard to observe. Can you elaborate on this?

Edited: I was wrong to believe that this gene had been linked to neurological diseases.

 

From the commentary:

what is known about TPPP? Hajdarevic et al. (2022) highlighted this gene’s high expression in brain, in particular in oligodendrocytes. This gene is also expressed more widely, including in ciliated lung and adipose cells, and only non-neurological diseases have systematically been linked to TPPP, namely lung disease, ulcerative colitis, Barrett's oesophagus, and haemorrhoidal disease (Mountjoy et al., 2021 https://doi.org/10.1038/s41588-021-00945-5.).

 

A problem with the microtubule network could lead to the axons of neurons dying. This is what happens in some neurological diseases. The effect depends on which neurons are affected.

I suspect it would be fairly obvious if this was occurring in ME/CFS. Maybe the resulting problem isn't serious enough to cause (much?) neurone death but enough to cause impaired function and an extended recovery period after stressors?

 

I don't think this study tells us anything useful. You can't do a GWAS study with ~400 participants.