ME Diagnosis and Genetic Testing and also, Gratitude

[JaimeS]

In case anyone else comes wandering in looking for information on SNP rs61750615, the link is correct, but you really have to hunt for the mention of it. It's solely discussed in the context of a table, and will not show up if you hit Ctrl-F:

Rare Variants with Known Clinical Associations:

Citation:

Dewey FE, Chen R, Cordero SP, Ormond KE, Caleshu C, et al. (2011) Phased Whole-Genome Genetic Risk in a
Family Quartet Using a Major Allele Reference Sequence. PLoS Genet 7(9): e1002280.
doi:10.1371/journal.pgen.1002280

 

[JaimeS]

@Valentijn

Congratulations, +/+ is the mildly faster, safer version You're one of the lucky 10% who get it.

Re: the CBS 699T - rs234706 - I found this today:

We found significant associations with LDL-C and total-cholesterol levels for a synonymous SNP (rs234706) in the cystathionine beta-synthase (CBS) gene (p = 1 × 10(-5) and adjusted-p = 0.013, respectively). Further, liver samples taken from 206 patients revealed that patients with the minor allele of rs234706 had significant dysregulation of CBS (p = 0.04).

n = thousands! This is probably significant. It is from here.

The article claims that +/+ CBS 699T has a statistically significant affect on cholesterol levels, LDL-C levels, triglycerides, and CBS function (homocysteine levels).

But perhaps the study isn't reputable. It looks good to me, but the statistical stuff goes over my head - I took it in college, but I made it by the skin of my teeth and never really 'got' it. (Also, my cholesterol and triglycerides are a wee bit low as a matter of fact, so there's an n=1 argument against a "strong correlation".)

It's actually a rather interesting article, and I recommend it to anyone with a scientific bent who's interested in the SNP thing - the diagram regarding how different metabolic disorders are connected to each other and to different genes is a thing of beauty.

-J

 

[Valentijn]

n = thousands! This is probably significant. It is from here.

The article claims that +/+ CBS 699T has a statistically significant affect on cholesterol levels, LDL-C levels, triglycerides, and CBS function (homocysteine levels).

It's hard to say. They used a different SNP (rs234705) with some linkage disequilibrium to determine rs234706 impact on gene function - but the linkage disequilibrium is not very strong, even though they call it "a perfect proxy". Reality is that the prevalence rates are fairly different in most ethnic groups who were tested for both.

Effect sizes were also pretty small ... +6% for total cholesterol, with a relatively high P-value for this sort of study. Usually with this many participants, there are much smaller p-values. For gene function, they're looking at a much smaller sample size (206), using only livers from people undergoing heart surgery (not the best cohort), and getting a barely significant result.

I'm also confused about their claims that rs234706 is significantly associated with HDL, LDL, and triglycerides. Their own Table IV contradicts that, with very high p-values for those three. It looks like they're also analyzing those values based on age and sex, which increases the odds of hitting a false positive somewhere. But they aren't showing those results, nor indicating how they broke them down by age, nor explaining how gender might reasonably be a factor. Did the age 36-40 males have slightly worse lab values + more of the "risk" allele? They aren't saying, and that makes me very suspicious.

According to the supplementary material, it looks like they tested about 650 SNPs ... but they never say that explicitly, and certainly not in the main paper, so I'm not sure if they corrected for that or not. Basically testing more SNPs = much bigger chance of a false positive. Which I think is why SNP association studies using a lot of SNPs and a lot of patients typically set extremely low p-values. Reading though their step-by-step methods, they mention "correcting" for age and sex, but not for the number of SNPs tested.

But perhaps the study isn't reputable. It looks good to me, but the statistical stuff goes over my head - I took it in college, but I made it by the skin of my teeth and never really 'got' it. (Also, my cholesterol and triglycerides are a wee bit low as a matter of fact, so there's an n=1 argument against a "strong correlation".)

That SNP would be having a tiny effect, at most ... so not particularly relevant, compared to other factors: other SNPs, diet, etc. And I'm not at all convinced by their statistics.

 

[JaimeS]

@Valentijn - thanks for the analysis.

Rs1799853 shows up as pretty rare on dbSNP. I have the TT genotype, which is associated with an average of 40% slower clearance of warfarin, which I thought could NOT be super-common. dbSNP shows the prevalence of this SNP to be extremely low. But it's not on the less than 10% list.

I could also be reading it wrong. *sigh* I haven't felt as intellectually lost in a long time as I have with genetics.

-J

 

[Valentijn]

But it's not on the less than 10% list.

Some are left off due to 23andMe and 1000 Genomes giving different rsIDs for the same location. Most of the time, that was resulting in listing non-rare SNPs, so we just had to chop them out if they didn't match up. Hence it does miss some.

 

[Butydoc]

@Valentijn - I understand. Your issue is with people's tendency to behave as though everything on planet earth, including Biology, behaves by the same rules. When I was teaching, I remember how frustrated I would get when my Chem students just wanted to memorize a series of steps and use that to solve every problem... and how some literally refused to believe me when I told them that wouldn't work.

@Sidereal - I grew up in a town by the ocean, so I'm sure I was exposed to mold on a day-to-day basis. I never lived in a house with a mold problem or spent a great deal of time in... moldy situations? Old abandoned shacks? (Or pricey NY flats that have been flooded with seawater - let's not discriminate.)

Speaking of maintaining a spirit of scientific inquiry, I do wonder about the mold stuff and some of the viral stuff too. I think the issue is that I don't know enough about blood tests' sensitivity levels, and I really, really feel like I want some control-group data. If you tested a bunch of people who feel fine for as many different pathogens as I have been tested for, what's the likelihood that they're going to come back 'infected', too, with something? Haven't we all been exposed to this stuff, and don't we have it in our bloodstreams too?

Come to think of it, in the context of my conversation about Von Willebrand's and @Gondwanaland's comment, maybe the simple reason why we ALL seem to have exaggerated reactions to medications is that many of us have low blood volume. This makes the medication more concentrated than it would otherwise be: more drug in the bloodstream per liter. Is this ridiculous and simplistic? It seems too simple to be right biologically, but it is true from the perspective of a chemistry person:

500-mg drug / 3 L vs 500-mg drug / 4 L

167-mg/L concentration vs 125-mg/L concentration.

Crowds the receptors - metabolites that may need to be recycled in the liver produced too quickly...

Might be part of the picture.

-J

Hi Jamie,

Most drugs are distributed among all the fluid spaces not just the blood volume. I would suspect that the reduced blood volume would be relatively small compared to the total body fluid volume hence probably not a huge change in drug concentration. In men, 65% of the body weight is water weight, whereas the blood volume is only 5 liters.

Best,
Gary

 

[JaimeS]

@Valentijn - whelp, I found an article that literally says the opposite regarding rs234607 (CBS C699T) and cholesterol, so I reckon you're correct about the crappy design of the study.

...although the study I was looking at was a small one, so that doesn't prove very much, either.

Also, this process inspired me to write a blog post regarding the nature of CBS C699T, and Valentjin, I came to the same conclusions as you did.

-J

 

[Valentijn]

@Valentijn - whelp, I found an article that literally says the opposite regarding rs234607 (CBS C699T) and cholesterol, so I reckon you're correct about the crappy design of the study.

...although the study I was looking at was a small one, so that doesn't prove very much, either.

Yeah, there's research going both ways, but stronger on one side than the other. Honestly, it's probably one of those SNPs which does bugger all - and if not, it's close enough to doing nothing that I don't think it's worth getting excited about either way

 

[jess100]

Hi everyone
I recently met a genomic doctor in Mexico-and wonder if she can be at all helpful to me since I already have a diagnosis of tick-borne and viruses. So would genomic information be any help to me? I was thinking no, because I already know the genesis of my illness-also I'm not sure how good she is or where she studied. Any advice?

 

[JaimeS]

@jess100 -- it's possible it might help you to get a (relatively) inexpensive test such as the one performed by 23andme. At the very least, you can see spots where you are more likely to be deficient in certain vitamins, or metabolic cycles that are more likely to be 'pushed' or slow. I don't think this information is super-useful to everyone, but in somebody with a weakened system, it might be a good plan to check out pre-existing sensitivities and potential deficiencies.

The thing about tick-borne illnesses is that most people appear to get over them.

Not medical advice or anything, but I'd get my immunoglobulins checked before I'd get the 23andme stuff done.

-J