FDA Orders 23andMe to Stop Providing Genetic Analysis.

[*GG*]

I think we should stop selling bathroom scales too.

I don't think they are calibrated accurately enough and might not be as valid as a scale should be or as a scale might be 50 years in the future.

How do we even know if a scale is accurate at all? Who is regulating the scales? Clearly they are dangerous and should be taken off the market immediately to protect the public from themselves.

And all those greedy little scale-sellers out there, just raking the money in, without any concern for whether or not their device is actually accurate.

Given the importance of weight on other serious health matters, like heart health and diabetes, obviously, scales should only be found in doctors' offices where people with the appropriate amount of training can be trusted to interpret the results and incite or not incite the appropriate degree of panic.

Where does this end???

It doesn't. That is why we need big govt to "protect" us from our stupid mistakes, because we all know that gov't doesn't make mistakes and/or hinder advances in science via regulations etc..

GG

 

[Valentijn]

To my recollection there were some who sent in two samples to 23andme,can't find that but haven't looked through all my files but did find where people had sent different results from two different companies.

Your recollection is incorrect. The only differences were between different companies, as you found in your files.

I thought the FDA cracked down because the validity and over interpretation of what the test means.

Your recollection is wrong again. The FDA only got involved because the genetic data was being interpreted. Their decision did not involve or impact upon the reporting of raw data. The FDA did not even claim that the interpretation was inaccurate. It was objected to solely because it was interpretation.

 

[taniaaust1]

Your recollection is incorrect. The only differences were between different companies, as you found in your files.

Your recollection is wrong again. The FDA only got involved because the genetic data was being interpreted. Their decision did not involve or impact upon the reporting of raw data. The FDA did not even claim that the interpretation was inaccurate. It was objected to solely because it was interpretation.

Its as Valentijn said. The raw data is recognised as being correct, hence why that is still available. They just didnt trust the interputations.. which is stupid when 23andME also had a five star scale thing in which showed how much such or how little each interputation was based on, so that people could make the decision for themselves what they wished to take in account and if one wanted to take into account an interputation which science could later prove wrong as it wasnt based on blind big studies but may of only been shown by a small study, that kind of thing only got a 1 star out of 5.

 

[taniaaust1]

.

Via this test I was shown to have an increased risk for breast cancer, did this make me run to the doc - er no......
I will continue on with the screening program and self check regularly, nothings changed.
Okay so if someone was shown to have a decreased risk - would that make them drop out of the screening program and stop self checking? er no..........

correct. I have breast cancer in my family and this test came out saying I was at a decreased risk with the genes they tested. It is relieving to know I dont have the common genes which cause it but will that make me think I wont get breast cancer... certainly not as it is clear there could be other genes involved in this..

Governments are forgetting that we are in the age of information and people nowdays are so much smarter over health things esp those who seek out sites like 23andME as they have independantly taken their own health issues into their hands.

 

[taniaaust1]

The confusion arose because they list the ethnicity of patients in the research which they used to interpret results. This was a somewhat stupid thing to do - a version of an SNP doesn't actually act differently in different races, but small studies will often get different results, and there are likely additional and unrelated factors having a bigger impact.

I know nothing about the case being talked of here but I do think that it is possible that a SNP could do different things in different races. SNPs can work together with other SNPs and some ethnicities may commonly have another SNP which can interact or rule out an issue a certain SNP is causing. (i havent read this but it would be a common sense thing if you think about it and how some SNPs can be better or worst in presence of other ones).

 

[taniaaust1]

That's the problem. It's not simple.

Barb

It really depends on what you want to do.. if you want to find out stuff about your methylation, its just a matter of putting the raw data into genetic genie which is quite easy. If ME/CFS patients can do that without issues, anyone can!

 

[Valentijn]

I know nothing about the case being talked of here but I do think that it is possible that a SNP could do different things in different races. SNPs can work together with other SNPs and some ethnicities may commonly have another SNP which can interact or rule out an issue a certain SNP is causing.

Exactly - it's still coming down to SNPs (or environment, diet, etc) not "ethnicity". Attributing differences to ethnicity is ultimately imprecise and sloppy. And in most cases, the results are the same for every ethnicity anyhow, so labeling ethnic groups is generally pointless and confusing.

 

[taniaaust1]

Exactly - it's still coming down to SNPs (or environment, diet, etc) not "ethnicity". Attributing differences to ethnicity is ultimately imprecise and sloppy. And in most cases, the results are the same for every ethnicity anyhow, so labeling ethnic groups is generally pointless and confusing.

Ethnicity is about inheritance and we inherit SNPs.

 

[Valentijn]

Ethnicity is about inheritance and we inherit SNPs.

Yes, but people in one ethnic group can inherit SNPs which are much more common in a different ethnic group, etc. Hence the ultimate factor is still SNPs, and ethnicity is just a rather vague and imprecise term in that context.

There is a great deal of variation in every ethnic group, and a great deal of overlap between very different ethnic groups. The ultimate factor is always the SNPs, and being a member of a specific ethnic group in no way guarantees that someone has the common SNPs for that group.

 

[WillowJ]

I know nothing about the case being talked of here but I do think that it is possible that a SNP could do different things in different races. SNPs can work together with other SNPs and some ethnicities may commonly have another SNP which can interact or rule out an issue a certain SNP is causing. (i havent read this but it would be a common sense thing if you think about it and how some SNPs can be better or worst in presence of other ones).

I think what tania is talking about is gene expression, which is indeed affected by other portions of the genome.

However, ethnicity would not play much of a role in differences caused by other SNPs affecting gene expression, as only "a small fraction of variance in allele frequencies" are involved in differences between populations.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1893020/

 

[nandixon]

Exactly - it's still coming down to SNPs (or environment, diet, etc) not "ethnicity". Attributing differences to ethnicity is ultimately imprecise and sloppy. And in most cases, the results are the same for every ethnicity anyhow, so labeling ethnic groups is generally pointless and confusing.

That's not correct according to my research. I've been meaning to write about this for a while and so have started a separate thread entitled: "Importance of ethnicity and allele frequency in determining disease-causing SNPs" (http://forums.phoenixrising.me/inde...cy-in-determining-disease-causing-snps.30710/)

Here's an excerpt from what I wrote:

"In summary, the scientific literature is clear that ethnicity can be a critical factor when attempting to narrow down what potential SNP or SNPs might cause or contribute to an illness, and that an ethnically appropriate genetic database should be used to avoid false negative results for SNPs that may be disease-causing/contributing and of lower frequency in one ethnic group but benign and of higher frequency in other ethnic groups.

There are dozens of research articles that have been published that emphasise this. Below are just a handful:"

 

[Valentijn]

That's not correct according to my research.

Are any of those alleles which supposedly show a different effect having a big impact? Sometimes there are different results between ethnicities in different studies or internally in a single study. But there are also often differences in different studies of the same ethnicity. When the effect size is pretty small, it's not unusual for one study to get a positive result, another gets the opposite result, and a third gets a null result.

 

[barbc56]

A blogger who writes about "mathematical curiosities" sent samples to 23andme two different times. Using stasticual analysis his results are below. But it's important to keep the first quote in mind.

The vast majority of DNA, more than 99.9% of the base pairs, is the same in all human beings. But with more than 3 billion base pairs in the human genome, even a 0.1% difference between two individuals comes to more than 3 million differences

.
http://roadontime.blogspot.com/2010/07/23andme-snps-so-nice-they-call-them.htm

Comparing the raw data, 85 SNPs were called differently by 23andMe in the two tests. Of these 85 SNPs, 73 were called as homozygous one time and heterozygous the other time. The remaining 12 were homozygous but opposite on the two occasions. (None of the differences were on the X, Y, or mitochondrial unpaired locations.)

This suggests a 23andMe error rate of 0.0074%, based on calls that differed on the two occasions.

The actual error rate is presumably somewhat higher than this, since some of the SNPs that were called one time but were no-calls the other time may also be incorrect, but I have no way of identifying those. It's also possible that there are a few systematic errors, which would be wrong nearly every time.

http://roadontime.blogspot.com/2010/05/my-spitting-image-23andme-error-rate.html

Here are his other blogs related to 23andme which go into greater detail about the statistics.

http://roadontime.blogspot.com/search/label/Sharing

 

[Valentijn]

@barbc56 - Another potential cause of different results can be that different cells sometimes have different mutations. It's not something I've looked into specifically, but I've come across mentions of it in cancers as well as in non-pathogenic circumstances. Though usually the differences are in different parts of the body, it's not always the case. But no idea how often it happens.

And I can confirm that 23andMe definitely does have systematic errors, where they report 100% of users as having a very rare genotype. I've only seen one, but someone else who noticed it commented on complete lack of responsiveness from 23andMe in fixing the error or even responding to it - maybe because that chip was designed and owned by a 3rd party, and it wasn't something they could do anything about themselves? It's not on the new chip now, so I guess it's not really an issue anymore

At any rate, 99.9936% accuracy is extremely impressive, especially for the price. Thanks for finding that data and sharing it!

 

[barbc56]

Hi @Valentijn

Where did you get the figure 99,39......% accuracy? From what I have read in the above mentioned articles the test is far from that accurate.

I'll check again as I think there was percentage mentioned.

Barb

 

[Valentijn]

Where did you get the figure 99,39......% accuracy? From what I have read in the above mentioned articles the test is far from that accurate.

It was what someone got from sending in two samples, as you quoted above.

 

[WillowJ]

That's not correct according to my research. I've been meaning to write about this for a while and so have started a separate thread entitled: "Importance of ethnicity and allele frequency in determining disease-causing SNPs" (http://forums.phoenixrising.me/inde...cy-in-determining-disease-causing-snps.30710/)

Here's an excerpt from what I wrote:

"In summary, the scientific literature is clear that ethnicity can be a critical factor when attempting to narrow down what potential SNP or SNPs might cause or contribute to an illness, and that an ethnically appropriate genetic database should be used to avoid false negative results for SNPs that may be disease-causing/contributing and of lower frequency in one ethnic group but benign and of higher frequency in other ethnic groups.

There are dozens of research articles that have been published that emphasise this. Below are just a handful:"

I agree that studies should not assume their results apply to all populations.

 

[Bob]

Anyone should feel free to to start a new thread about this...
23andMe have got new permissions from the FDA... I'm not certain what the exact details are...

DNA screening kit given the go-ahead by US regulator
A Californian start-up will be allowed to advertise a mail order DNA test that screens for a rare genetic condition, after a U-turn by the US regulator.
http://www.bbc.co.uk/news/technology-31552029

 

[Valentijn]

23andMe have got new permissions from the FDA... I'm not certain what the exact details are...

Yeah, the initial ban on genetic interpretation was pretty stupid. Yes, there are a lot of SNPs which have very little or likely no impact, and customers have trouble understanding it, but there are also SNPs which are 100% pathogenic. It never made any sense to ban info regarding those proven pathogenic SNPs.

 

[wdb]

Anyone should feel free to to start a new thread about this...
23andMe have got new permissions from the FDA... I'm not certain what the exact details are...

DNA screening kit given the go-ahead by US regulator
A Californian start-up will be allowed to advertise a mail order DNA test that screens for a rare genetic condition, after a U-turn by the US regulator.
http://www.bbc.co.uk/news/technology-31552029

It sounds like a step in that direction but a complete U-turn yet

Last night, the Food and Drug Administration approved a test made by 23andMe, the Mountain View, Calif.-based personal genetics company, for a gene that can cause a rare disorder called Bloom Syndrome, which causes short stature and a heightened risk of cancer.
...
The FDA is not going back on its 2013 decision to stop allowing 23andMe to sell its personal genetics tests directly to consumers. Right now, this decision applies mainly to the Bloom test.

http://www.forbes.com/sites/matthew...da-approval-means-for-the-future-of-genomics/