23andme and privacy

[karendo12]

Hi,

I would like to order a 23andme kit, as I understand that many of you may have.
My concern is that genetic information that is tied to me can be used against me in the future . For instance, if a hacker gets hold of the information. Having genes predisposing someone to a severe illness or dementia might result in job or other discrimination. In the near future everyone will have their genome sequenced and no one knows what the consequences of having one's genes exposed will be.
Does anyone know how one can order such a kit anonymously? I do not live in the U.S. and 23andme does not send to P.O. boxes outside the U.S.

Thank you in advance.

 

[Sea]

You don't have to give them your real name and you can use a friend's address for posting.

 

[Eeyore]

Technically you are required to give them your real name. This is probably not enforceable though.

There are alternatives to 23andme that don't store your data long term - such as https://www.genebygene.com/pages/research#

You can do the array test for about $200. I believe they send the results to your doctor who can then share them with you.

They don't test the exact same snps as 23andme, but they're going to both test most of the more interesting snps. The 23andme v4 chip is cheaper/smaller anyways.

You can also try ancestry.com - I'm not sure if they require a real name. Their chip may be the same as genebygene - not sure - but again, it's going to test a lot of snps and probably include the important ones for most people.

 

[Valentijn]

There are alternatives to 23andme that don't store your data long term - such as https://www.genebygene.com/pages/research#

You can do the array test for about $200. I believe they send the results to your doctor who can then share them with you.

Wow, that one is looking very nice. They target SNPs with a frequency of 5% and less, plus various missense mutations. So a lot less background spam than we get from 23andMe, yet they still give data for over 700,000 SNPs - more than 23andMe currently. And the price tag is quite reasonable as well.

 

[Eeyore]

@Valentijn

Yes, I looked around a bit at the more reasonably priced ones. I ended up going with 23andme for my first round because we did it as a family, and having family data can be very useful. I was able to phase my genome against the other 3 members of my immediate family (including both parents) and a reference panel of 2500+ haplotypes using impute2. This allows much more accurate imputation for missing SNPs and helps me weed out a lot of Mendel errors, and when something is weird, I can make sure one of my parents has it too.

They seem to always have no calls on important SNPs. e.g. APOE and BRCA1/2 seem, even after opting in, often give no-call results. I had mine for APOE since my endocrinologist routinely runs it as part of his cardiovascular management approach (he's extremely forward thinking and focuses on preventive care a great deal). I was able to accurately impute my entire family's genotype though, even though 3 were missing - turns out we are all e3/e3 except my dad who is e2/e3 - that fits since we have no family history of alzheimers at all. With 23andme, you can probably impute the vast majority of SNP's that are missing except the extremely rare ones, where the reference panel is probably not large enough to accurately gauge D'.

One SNP of mine that was a no-call and important took me about a week to conclusively figure out.

I'm still looking at whole-exome sequencing or whole-genome sequencing, but haven't gone for it yet. I think a good SNP array tells you a lot right now in the meantime.

You'd be amazed how many Mendel errors some reports have - mine wasn't terrible (maybe 10?) but my sister's had many times that. I think I used plink for that, it's pretty easy when you create a ped/bed file to output a list of mendel errors if you have both parents. Of course, that doesn't necessarily tell you where the error actually is.

 

[karendo12]

Thank you all for your replies.
I have been suffering from CFS for many years and am looking into anything that might give me a clue as to what might be wrong. There are methylation defects in my family and perhaps that is one of my problems as well.

 

[Eeyore]

Personally, I doubt that methylation is the root of the problem in ME. I think, however, that it can perhaps improve some symptoms to some degree - the science is still not there to say for sure.

I actually started taking B12 because I was found to have a functional deficiency (elevated MMA). This was before I even knew what methylation was in the context in which it is now used. I noticed a huge beneficial effect. My father does also - my mom doesn't notice anything at all. Placebo? Maybe - but considering I can document a deficiency, I think it less likely.

Later on I was tested and found to have hyperhomocysteinemia, so my doctor wanted me to take folate. It didn't help. He then ordered a test (he's not an ME doc, and he wasn't trying to treat ME) for the MTHFR mutation. He told me that I wasn't able to methylate folate efficiently and should take methylfolate, and wrote me a prescription for cerefolin NAC. It brought my homocysteine down.

Years later, I read about RVK's methylation theories, and considering that my docs got there on their own, at least in part, years ahead based on my bloodwork suggests to me that there might be something there. I continue to take cerefolin primarily because it reduces cardiac risk (probably - mechanistically one would predict it, although I'm not sure any study has shown it, but it seems better to me to follow the more likely, if unproven, course than to sit around and wait for science, knowing that I'm probably causing myself harm by waiting).

I do think methylation has been overblown as the main cause of ME or the first line treatment. However, I think it may be a good supportive measure in general (maybe). Hopefully time and science will give us better answers. Unfortunately, for now, we must make decisions based on what we know now, which remains very limited. If I can't get proof either way, I'll take the course that seems, even if only slightly, more likely beneficial/not harmful.

My doc also believes that the evidence currently out there suggests lowering homocysteine is beneficial in prevention of heart disease. He does warn that too much folate has been associated with colorectal cancer though. We talk through it together and come to a reasonable approach that we think takes into account all the information that is currently available to us. Don't fall into the trap though of assuming that supplements are harmless. If they are potent enough to help, they are potent enough to cause negative side effects.

 

[karendo12]

Thanks, Eeyore.
I'm glad the B12 helped you.
You are right about supplements being potentially harmful. Regarding the risk involved with high folate, I thought that was only true with the regular form of folic acid. Have you heard the same about methylfolate?

 

[Sea]

They seem to always have no calls on important SNPs. e.g. APOE and BRCA1/2 seem, even after opting in, often give no-call results. I had mine for APOE since my endocrinologist routinely runs it as part of his cardiovascular management approach (he's extremely forward thinking and focuses on preventive care a great deal). I was able to accurately impute my entire family's genotype though, even though 3 were missing - turns out we are all e3/e3 except my dad who is e2/e3 - that fits since we have no family history of alzheimers at all. With 23andme, you can probably impute the vast majority of SNP's that are missing except the extremely rare ones, where the reference panel is probably not large enough to accurately gauge D'.

One SNP of mine that was a no-call and important took me about a week to conclusively figure out.

I read about the no-calls on the 23andme website and remember it said it was due to inaccuracies on particular runs. So everyone who had their test done on the same run would receive the same no-calls. Having the test redone on another run would likely result in different no-calls. In some circumstances 23andme have offered to rerun some tests.

I was given my APOE status e3/e4 by 23andme, no no-calls on mine.

I'm interested in imputing snps, is this only possible when you have your parents results as well? I remember looking at a programme for this once but I couldn't work it out and I don't remember which one it was.

Later on I was tested and found to have hyperhomocysteinemia, so my doctor wanted me to take folate. It didn't help. He then ordered a test (he's not an ME doc, and he wasn't trying to treat ME) for the MTHFR mutation. He told me that I wasn't able to methylate folate efficiently and should take methylfolate, and wrote me a prescription for cerefolin NAC. It brought my homocysteine down.

MTHFR snps influence the amount of folate required, but I am yet to find studies that indicate that MTHFR snps influence the ability to process folic acid even though it is common opinion that they do. I have asked on a couple of MTHFR sites if anyone can point me to such studies but sadly have instead been treated rudely for asking the question. I have found studies that indicate some people cannot process folic acid but they haven't been related to MTHFR genotype. One would expect the cases of neural tube defects to rise (or at least remain unchanged) in infants of MTHFR affected mothers who are exposed to the mandated folic acid supplements in our food, however that does not seem to be the case. Overall the rates have fallen regardless of MTHFR status.

I have however found studies that indicate DHFR may be involved in the inability to process folic acid.
Have you looked at your DHFR results on 23andme?
I only have one MTHFR A1298C snp but have several DHFR ones and have found benefit from supplementing with methylfolate.

 

[Eeyore]

It works better if you have your parents (or even just 1 parent). That's because genomic data you get is unphased - you don't know which is on which chromosome. Because things that are on the same chromosome are passed together, you can figure out which parent particular alleles came from.

When imputing, you actually impute against phased haplotypes - e.g. things on one chromosome. These tend to match up between various people pretty well, and with a large enough pool, you can frankenstein yourself a whole genome.

You can still do phasing and imputation w/o parents, but the accuracy is not as good. Others who are related to you are helpful too - siblings, cousins, aunts, uncles, etc. Even inlaws can be helpful if you also have a person who is related to both them and you (e.g. if you have your mom's sister's son, your' mom's sister's husband could be useful, as it helps to determine which of the son's genes are on each chromosome).

@karendo12 - I don't know if there are equal risks with methylfolate. I don't have a molecular understanding of why there are risks with folate in those cases.