H1 or J1 Haplogroups and rs3928306 (POLL)

[Valentijn]

I finally figured out a stupidly simple (and extremely slow) way to compare all of the 23andMe data we have for 12 ME patients to 12 controls using Excel. I then calculated how big the differences are for each SNP, and ranked the SNPs starting with the biggest difference between ME patients and the controls.

In the process I did the same for Mitochondrial DNA, which is passed directly from mother to child over the generations. In the MT-DNA, we had an allele (A) for one SNP ( rs3928306 ) which was present in 7 ME patients (58.3%) and 0 controls (0%). It's a known marker for the H1 maternal haplogroup, though also appears in the J1 haplogroup, and probably some others as well. The A allele has 16.5% prevalence in the general population.

So I'm curious to know what people's haplogroups are, and which allele they have for rs3928306. There's a similar haplotype poll in this subforum, but it's far less focused. Some additional haplogroup options are included based on what's somewhat common in our results so far, and closely related haplogroups.

Please don't answer the poll yet unless you know either your maternal haplogroup or your rs3928306 status. This poll is also intended for ME/CFS patients only, with CCC- or ICC-defined ME and Post-Exertional Malaise.

 

[Valentijn]

I also made a list of the haplogroups of all of the ME patients and controls, to make sure I hadn't randomly selected a group of controls who had a very different ethnic origin compared to the patients.

11 of 12 ME patients showed very European origins, though with a bit of NW Africa, the Near East, and North Asia being somewhat possible (H16, U4a1, J1c, J1c2a2, H1b1, H3y, H1, I2, H1C1, H2a1, H2a1c), and 1 ME patient likely from the Middle East or very SE Europe origins (J2a1a1a2).

1 control also had some likely Middle East or very SE Europe origins (J2a1a1), 1 with likely Indonesia or other SE Asia origins (F1a4), one with S Asia origins, one with Arabian origins (H18), and 8 with origins from around Europe or very near by.

One thing I'm considering now is finding controls with ancestry more closely matching our ME patients, as that might help to minimize the appearance of irrelevant variations which are due solely to normal ethnic variation. Though then I have to decide if I want controls who match our haplogroups, or just match our general areas of origin. I'm hesitant to deliberately match haplogroups, in case haplogroup is relevant to the disease in any manner. And it might be overkill anyhow.

 

[Valentijn]

I've gone ahead and matched up the full 23andMe data I have for 12 ME patients with 12 controls who have the same or very similar haplogroups. Basically it's the only reliable way to get controls who are very likely matched by ethnicity. While I was at it, I found female controls to make the X chromosome a bit easier to deal with

I decided to go with this approach because I think having well-matched controls offers several advantages, and only one unlikely disadvantage. While it is possible that maternal haplogroup is associated with the risk of developing ME/CFS, it would have to be a relatively weak risk. Together, the two haplogroups have an average of 10-20% prevalence throughout Europe, and is much higher in some places. Whereas ME/CFS is much rarer.

And by having nicely matched controls, it's easier to see if rarities are truly unusual, or just an expected result of having a certain ethnicity. So having two Asian controls could have been bringing up "rare" results that are actually quite common among Asians, but would have made actually rare results in a European sample look less rare than they are when compared side by side in a graph.

So my new graphs still have 12 patients and controls, with a thick line down the middle, but now P1 (patient 1) and C1 (control 1) are matched by sex and haplogroup. Similarly P2 and C2 are matched up, and P3 and C3, etc. So if there are any rare results purely due to ethnicity, it should be apparent just by glancing at the graph, as the rare results will match up spatially on the patient and control sides of the graph.

I'm looking at genes related to CoQ10 currently, and so far it's looking quite good. There seems to be few rare results shared by matched patients and controls, except in the mitochondrial DNA (MT-DNA) where similarities are to be expected, since that's how haplogroup is determined in the first place.

But something interesting I noticed in the MT-DNA is that the "no-call" are very much the same for patients and their matched controls. That is, if both haplogroup H1 patients have "no call" for an SNP, both H1 controls also have "no call" for that SNP. Hence it seems that no-call results aren't completely random, though I couldn't guess as to what it might mean - or if that significance would extend to the normal chromosomes.

 

[Valentijn]

As an update, now that I have 33 sets of results for ME patients:

12 out of 33 have rs3928306 A, or 36.4%. The normal prevalence in Europeans, based on a sample of 404, is 24.3%, and it's lower in almost all other ethnic groups.

So that's a 50% increase of that allele in ME patients compared to the general population, albeit with a small sample being used for comparison.

 

[Gondwanaland]

This poll is also intended for ME/CFS patients only, with CCC- or ICC-defined ME and Post-Exertional Malaise.

Sorry, I voted before reading that. I found this poll so exciting that I went right ahead.

AFAIK there is no diagnosis for CFS in my country, but I do have PEM that can last a few days.

My haplogroup is D1 and my SNP is A. I am 95.5% European.

My husband's haplogroup is U5a1a1 and SNP G. He has exercise intolerance and several neuropathies. He is 100% European.

 

[Sidereal]

My haplogroup is H11a. Does that fall under H1 or Other H in your poll?

 

[Valentijn]

My haplogroup is H11a. Does that fall under H1 or Other H in your poll?

That would be H-eleven, not H-one-one. So other H.

 

[Valentijn]

There wasn't really any info on where exons were for that gene, MT-RNR2, or which variants resulted in missense mutations. So by looking up the protein, I was able to figure out what some chunks of the SNPs should look like, and search for it in the map view of the MT-DNA on dbSNP.

Long story short, the coding part of MT-RNR2 starts at position 2,633 and ends at 2,704. But rs3928306 is at position 3,010 so it definitely isn't a coding SNP. Though it might be linked to another SNP which is a coding one, or it might be regulatory for the gene.

Few SNPs are identified with rsID's on the MT-RNR2 gene, and 23andMe only tests four of them, three of which can result in missense mutations:
i4001264 C->T results in Proline->Serine
i3002040 T->A results in no change
rs28602676 G->A results in Valine->Methionine
i4001262 G->A results in Alanine->Threonine

None of these are particularly likely to cause problems, as they're mutating into somewhat similar proteins. The BLOSUM62 scores are -1, 4, 1, and 0, respectively. Higher values are least likely to cause issues, and the values can go down to -4.