Rare Coenzyme Q10 gene variations in ME patients

[Leopardtail]

Agreed, which is why I'm looking at entire groups of genes together which are involved in a similar function, instead of just looking at single SNPs or genes in isolation. Though I do look at single rare SNPs we have in common, and want to do some more efficient sorting of rare gene results too for the entire file.

Yeah, that helps! But I don't much about these genes precisely, just what I mentioned above: they involve ubiquinol and the electron transport chain, which seems to be required for cellular respiration and/or energy.

Yes, that's something I want to work on. I have 3-4 or more newer sets of 23andMe data from ME patients, and tons of older controls sets, but I can't combine it with the older version I have for the 12 ME patients shown above. My laptop simply doesn't have the CPU power, memory, hard drive space, or good software to do it.

If I separate out alleles from the genotype to make manipulating the data easier, Excel pretty much bursts into flames. It either hangs for 10 minutes to apply a formula to a column, or it flat out refuses to perform the operation and then can't be saved at all until I reboot.

My laptop is falling apart a bit anyhow, so my fiance is working on getting me a very fast PC assembled. But it's costing a fair bit of cash, which we currently have to pay toward medical expenses while waiting for the insurance company to reimburse us (or not). If they decide not to, my parents will give us the cash, but until that decision is made by the insurance company, I have to wait But there will be a better PC in the next month or two ... 8 core 3.5GHz CPU, a good-sized solid state hard drive, and 16GB memory. It'll be fast

Yes, that's why I left that one in, even though it's not quite rare enough to meet my usual standards. "CT" at rs3822662 has a calculated prevalence of 12.5% in the general population, but we have it at 75% and the controls have it at 8.3%. Nothing is known about the SNP, so maybe it's just random coincidence. But then again, maybe not

@Valentijn

You will know half of this, but including info for all who read.

The role of ubiquionol in the electron transport chain (ETC) is very simple it's one of the links in the chain. In order to make a new ATP synthase (the physical enzyme that implements the ETC) you need a ubiquinol molecule - each mitochondria needs very many of that enzyme hence very many Ubiquinol molecules. In simple terms the ETC is nothing more than a series of molecules that accept, then pass on electrons it's like a uni-directional 'copper wire'. At the end of that 'wire' ATP is made. Most cells also need many mitochondria.

Hence requirement = many (per mito) x many (#mitos)

Each mito should have a functional life of six weeks then divide into younger fresher mitos:

Hence requirement = many (per mito) x many (#mitos) x few (semi-frequent division)

Next we have to account for the 'oxidative stress' in PWME - ubiquinol 'disarms' free radicals but becomes the useless (for the ETC) ubiquinone meaning
Hence requirement = many (per mito) x many (#mitos) x few (semi-frequent division) x very many (oxidative loss)

Add in poor production of CoQ10 and you end up with
Mito status Status = Production (low) / Requirement (very high)

This seems likely to result in very few, or very geriatric mitos since you can't make functional mitos without the 'right CoQ10' = ubiquiniol.

 

[SDSue]

@Valentijn Such interesting stuff - thanks for putting it all together. I know very little except that NDUF7 rs7258846 is the known SNP associated with Leigh Syndrome, an early onset severe form of MD.

Any idea what to make of my NDUF7's? It doesn't seem to paint a pretty picture.

 

[Leopardtail]

Agreed, which is why I'm looking at entire groups of genes together which are involved in a similar function, instead of just looking at single SNPs or genes in isolation. Though I do look at single rare SNPs we have in common, and want to do some more efficient sorting of rare gene results too for the entire file.

Yeah, that helps! But I don't much about these genes precisely, just what I mentioned above: they involve ubiquinol and the electron transport chain, which seems to be required for cellular respiration and/or energy.

Yes, that's something I want to work on. I have 3-4 or more newer sets of 23andMe data from ME patients, and tons of older controls sets, but I can't combine it with the older version I have for the 12 ME patients shown above. My laptop simply doesn't have the CPU power, memory, hard drive space, or good software to do it.

If I separate out alleles from the genotype to make manipulating the data easier, Excel pretty much bursts into flames. It either hangs for 10 minutes to apply a formula to a column, or it flat out refuses to perform the operation and then can't be saved at all until I reboot.

My laptop is falling apart a bit anyhow, so my fiance is working on getting me a very fast PC assembled. But it's costing a fair bit of cash, which we currently have to pay toward medical expenses while waiting for the insurance company to reimburse us (or not). If they decide not to, my parents will give us the cash, but until that decision is made by the insurance company, I have to wait But there will be a better PC in the next month or two ... 8 core 3.5GHz CPU, a good-sized solid state hard drive, and 16GB memory. It'll be fast

Yes, that's why I left that one in, even though it's not quite rare enough to meet my usual standards. "CT" at rs3822662 has a calculated prevalence of 12.5% in the general population, but we have it at 75% and the controls have it at 8.3%. Nothing is known about the SNP, so maybe it's just random coincidence. But then again, maybe not

is this a simple matter of pasting data from an old spreadsheet into a new one?

 

[Valentijn]

is this a simple matter of pasting data from an old spreadsheet into a new one?

Nope. The problem is that the older set of data has 960,000 SNPs, and the newer set has 600,000+ SNPs. These two sets need to be merged, but Excel hits the limit of its rows well before 1,600,000. So I can't even get both sets into the same sheet as would be needed to merge them. Which is annoying, because they'd definitely be under the limit once merged.

What I really need to do is switch to using a database instead of a spread sheet. That will also make it easier to automate the calculation of expected genotype prevalence, tagging each SNP with the name of the gene it's on, tagging known missense and/or pathogenic mutations, converting "i" numbers to "rs" numbers, etc etc.

 

[Valentijn]

@Valentijn Such interesting stuff - thanks for putting it all together. I know very little except that NDUF7 rs7258846 is the known SNP associated with Leigh Syndrome, an early onset severe form of MD.

Any idea what to make of my NDUF7's? It doesn't seem to paint a pretty picture.

All versions of those SNPs are very common. And 4 of them are tightly linked, so having all 4 is no more relevant than just having one. Ironically, those linked ones include 3 "red" results and one "green" result, so they're essentially contradicting themselves by flagging 3 of them but not the 4th. Or they're deliberately raising a false alarm, as everyone will have to be red for either one or the other three of the SNPs.

 

[Leopardtail]

@Valentijn Such interesting stuff - thanks for putting it all together. I know very little except that NDUF7 rs7258846 is the known SNP associated with Leigh Syndrome, an early onset severe form of MD.

Any idea what to make of my NDUF7's? It doesn't seem to paint a pretty picture.
View attachment 7954

I don't any data on that on V's report sue - I'm confused!

 

[Leopardtail]

Nope. The problem is that the older set of data has 960,000 SNPs, and the newer set has 600,000+ SNPs. These two sets need to be merged, but Excel hits the limit of its rows well before 1,600,000. So I can't even get both sets into the same sheet as would be needed to merge them. Which is annoying, because they'd definitely be under the limit once merged.

What I really need to do is switch to using a database instead of a spread sheet. That will also make it easier to automate the calculation of expected genotype prevalence, tagging each SNP with the name of the gene it's on, tagging known missense and/or pathogenic mutations, converting "i" numbers to "rs" numbers, etc etc.

It's definitely a job for a database, but would likely break Access as well. I suspect that's going to be a job for mysql (faster) or postgreSQL (more robust with complex queries).

How IT literate is your fiance when it comes to complex databases?

 

[Valentijn]

I don't any data on that on V's report sue - I'm confused!

That's because I'm only listing rare SNPs. The ones in SDSue's report have alleles with 38% and 45% prevalence in the general population. So approximately 15% of humanity is also homozygous for one set, and 20% are homozygous for the other set.

 

[Valentijn]

It's definitely a job for a database, but would likely break Access as well. I suspect that's going to be a job for mysql (faster) or postgreSQL (more robust with complex queries).

How IT literate is your fiance when it comes to complex databases?

Extremely He also has a dim view of Access, as it's actually pretty much a spreadsheet as well and not a proper database program. The new PC is planned to have Linux installed as we both hate Windows 7 and 8, and XP is too old to work well with the 8 core processor. Though it sounds like Windows in general doesn't utilize multi-core processors very effectively anyhow.

 

[Leopardtail]

Nope. The problem is that the older set of data has 960,000 SNPs, and the newer set has 600,000+ SNPs. These two sets need to be merged, but Excel hits the limit of its rows well before 1,600,000. So I can't even get both sets into the same sheet as would be needed to merge them. Which is annoying, because they'd definitely be under the limit once merged.

What I really need to do is switch to using a database instead of a spread sheet. That will also make it easier to automate the calculation of expected genotype prevalence, tagging each SNP with the name of the gene it's on, tagging known missense and/or pathogenic mutations, converting "i" numbers to "rs" numbers, etc etc.

so does the original data set have un-needed SNPs that could be removed (thus producing a clean data set)?

Are they for different populations of patients without intersection?

 

[Leopardtail]

Extremely He also has a dim view of Access, as it's actually pretty much a spreadsheet as well and not a proper database program. The new PC is planned to have Linux installed as we both hate Windows 7 and 8, and XP is too old to work well with the 8 core processor. Though it sounds like Windows in general doesn't utilize multi-core processors very effectively anyhow.

I had real issues with 64 bit linux, be careful there unless you are much better at LINUX than me

 

[adreno]

@Leopardtail

So you think PWMEs would benefit more from supplementing ubiquinol, rather than ubiquinone, even though they should interconvert?

 

[heapsreal]

Are people really seeing a difference between ubiquinone and ubiquinol?

I have read a few things of late that it doesn't make a difference and more a marketing ploy to sell more expensive q10.

I have only used ubiquinone so I can't give an opinion either way other then since using high dose of ubiquinone, that it has helped.

 

[Valentijn]

so does the original data set have un-needed SNPs that could be removed (thus producing a clean data set)?

Are they for different populations of patients without intersection?

I've actually already tried removing entries where all 12 ME patients and controls have exactly the same results, and that takes us down from 960,000 rows to 842,000 rows. But it's still too many for Excel when adding the newer 23andMe results, especially if I want to be able to manipulate data after adding additional patients and/or controls. That approach would also risk missing rare results from new patients whose data is added, if those SNPs have been filtered out.

The fiance thinks mysql and postgresql wouldn't be able to handle the number of columns which will be needed in a proper database. He wants us to use the Apache Cassandra database management system, which should be able to handle it. And he's quite good with Linux too, so shouldn't have any problem getting me set up

 

[Valentijn]

so does the original data set have un-needed SNPs that could be removed (thus producing a clean data set)?

Actually, maybe I can merge approximately half of the new and old 23andMe chips' data at a time, pending my new system being set up. It might work - if having the extra patients and controls doesn't kill Excel once the merged halves are then merged together

 

[Leopardtail]

@Leopardtail

So you think PWMEs would benefit more from supplementing ubiquinol, rather than ubiquinone, even though they should interconvert?

There is not solid science on that yet. What's certain is that during mito division we need ubiqinol, and ATP must be available to make it from ubiquinone.

The strategy I am trialling at the moment is mixed supplementation with 200mg ..ol and 200mg ...one at different times totalling 400mg / per day. The bit of Q10 I understand well at the moment is the ATP synthase part, with heavy fatigue I strongly suspect Ubiquinol is needed, but it's vastly more expensive and harder to get (you can buy powdered ubiquinone). I have still to do work on the other functions so not sure how things lie there yet. I did find that link on Cholesterol and CQ10 though. My suspicion is always the body makes different use of different versions of things.

I have a lot of ME 'triggers' at the moment, so hard to know for certain how well it's working.

 

[Leopardtail]

I've actually already tried removing entries where all 12 ME patients and controls have exactly the same results, and that takes us down from 960,000 rows to 842,000 rows. But it's still too many for Excel when adding the newer 23andMe results, especially if I want to be able to manipulate data after adding additional patients and/or controls. That approach would also risk missing rare results from new patients whose data is added, if those SNPs have been filtered out.

The fiance thinks mysql and postgresql wouldn't be able to handle the number of columns which will be needed in a proper database. He wants us to use the Apache Cassandra database management system, which should be able to handle it. And he's quite good with Linux too, so shouldn't have any problem getting me set up

Why does he want to use so many columns rather than separate the results into a separate line data table? That would lead to far more options for reporting, quizing, and stats.

 

[Leopardtail]

Are people really seeing a difference between ubiquinone and ubiquinol?

I have read a few things of late that it doesn't make a difference and more a marketing ploy to sell more expensive q10.

I have only used ubiquinone so I can't give an opinion either way other then since using high dose of ubiquinone, that it has helped.

there is science to indicate differences, but not impact on ME patients. It's like the comparison between r5p and riboflavin.

 

[Leopardtail]

I've actually already tried removing entries where all 12 ME patients and controls have exactly the same results, and that takes us down from 960,000 rows to 842,000 rows. But it's still too many for Excel when adding the newer 23andMe results, especially if I want to be able to manipulate data after adding additional patients and/or controls. That approach would also risk missing rare results from new patients whose data is added, if those SNPs have been filtered out.

The fiance thinks mysql and postgresql wouldn't be able to handle the number of columns which will be needed in a proper database. He wants us to use the Apache Cassandra database management system, which should be able to handle it. And he's quite good with Linux too, so shouldn't have any problem getting me set up

not used cassandra so look forward to hearing how you get on.

 

[SDSue]

Thanks, @Valentijn - So good to know!