Ask HN: Is 23andMe DNA testing worth it?
See my friends getting tested and what interest me is the genetic conditions you may be more pre disposition for. But having a private company with my DNA is privacies of all privacies out the window. Is it worth? General thoughts?
86 comments
[ 2.6 ms ] story [ 31.6 ms ] threadIt sounds like his doctor is quoting this study, which does indeed suggest a 40% false positive rate: https://www.nature.com/articles/gim201838
Basically, almost all (~99.9%) of the results of tests like 23andme (SNP arrays) are correct, but because an interesting/scary variant/mutation is about as rare in the population as the test result being wrong for a given site, somewhere around half of those scary results are false positive. If 1000 people are tested for a mutation with this method, 10 will be false positive. The real frequency of mutations vary by site - some will be close to 50%, others at 10e-9, depending on how strongly that variant is selected against, so the most harmful ones are very rare. If you get a result from 23andme that your doctor thinks is worth retesting, it's probably because it's a harmful mutation, which is rare, and ergo there's a fair chance the 23andme result is wrong. So 23andme is almost always right, except when the result might affect your health in a very serious way. I should point out that for late-onset diseases like cancer, heart disease, alzheimer's, etc., there's little natural selection, so pathogenic variants are more common, making it more likely that the 23andme result is correct.
It's a simple bayes rule result. It would not be hard for 23andme to give you an estimate of how likely the result is to be a false positive, but maybe there's some regulatory barrier in the way?
How is error rate being defined in this case?
I took your explanation to mean: Cancer is common, so the genetic blueprint for Cancer risk factors should also be common. We have a higher number of samples to verify against which increases the detection accuracy.
For less common diseases, we have a smaller pool to verify against, so our blueprint is less accurate. Lower detection accuracy.
Disease X is very bad and strikes around age 30, so it has been selected against and is very rare. Only 1/10,000 people have the gene that causes it. Disease Y is bad, but it strikes when you're 70 (Alzheimer's let's say), so it is not strongly selected against and is more common, with 1/100 people having it.
The technology 23andMe uses is wrong about 1/10,000 times, just randomly, by chance. This is true for every place in the genome it tests.
Of 10,000 people, only 2 will test positive for disease X while 101 will test positive for disease Y. The 2 people who test positive for disease X will be the 1 person who actually has it, and 1 false positive, so 50% of the people who tested positive are actually at risk. Of the 101 people who test positive for disease Y, 100 will be at risk and 1 will be a false positive, so 99% of people who test positive are actually at risk.
Incidentally, this is why testing populations at very low risk for disease is generally counterproductive. The false positive rate stays the same but as the true positive rate gets very low, one tends to cause more harm than good.
> The technology 23andMe uses is wrong about 1/10,000 times
Is 23andMe's tech considered state-of-the-art? I'm assuming there are ways to reduce both the false positive and false negative rate and I'm curious how other labs/tech stack up against these rates.
If anyone is interested in a 23andme $40 settlement coupon from a class action I was part of, email me (email in profile), and I will provide a coupon code (I have three available).
Disclaimer: 23andme customer, genome is public as part of Harvard’s Personal Genome Project
EDIT: @monocasa: The future is what we make of it.
jstarfish: As a parent, I have to make what I consider to be the best choices for them until they're able to make their own choices. As an educated parent, I believe I made the appropriate choice. Everyone's DNA will be sequences eventually.
Of course, we've also seen them reconstructing likely genetics from family members on these sites, so it doesn't really matter whether you personally opt in or not.
I know the response is "Well I ani't killing/raping so it doesn't matter."
A lot of people have been sent to jail on false positives on dna results. It takes to much time to actually compare a large chunk of dna so they select certain parts to compare it to and if you fall into the unlucky percentage then you are screwed.
HOWEVER
If any of your family members have done a test like this, the cat is already out of the bag. My parents have done genetic tests, as have several of my cousins. So in effect, my DNA is already on file.
Seems kind of scary to me. What happens if I'm denied insurance, or a job because of some risk factor I have. Or even my kids?
Or even discover a way to make an drug even more addictive.
1: https://www.cnbc.com/2018/08/04/4--risks-consumer-face-with-...
The genetic side didn't have anything meaningful (but maybe because whatever markers it tests for, I didn't have any risk factors).
I also loaded the data into Prometheus, but all it gives you is academic journal articles that you'd have to sift through and do your own interpretations.
They recently partnered with a genetics company to share data - at that point, I opted out of data sharing since I was unsure how my data was exactly going to be used.
Also watch for discount deals - don't pay full price - you can get 30% off the kits during sales on Amazon.
tl;dr: do ancestry if its on discount, unsure about genetics.
I suppose if the reports showed _something_ unusual, it could have been worth the effort, so being disappointed is actually a good outcome.
They also tend to market heavy discounts (I've seen 50% some years) during Thanksgiving/Christmas, both of which are right around the corner.
The result of cheap genetic testing when family is in town must make for some awkward holiday dinner conversations...
Your genes indicate you’re likely predisposed to some diseases X, Y, and Z. There is no vaccine from any of these, but you can reduce likelihood of those diseases if you exercise regularly, but not too vigorously, mind your diet, reduce stress, drink rarely, and have satisfying social life. Also remember your annual medical exam, and consult a medical professional when things feel wrong.
This is true for nearly any person as I understand.
(A sample of the questions that I might be able to answer with actual specifics based on an analysis of my genes)
All of them need to be up to par.
For a quick explanation to those unititiated in auto physics: bald tires on front = increases understeer. Bald tires on rear = increases oversteer. Understeer = car pushes straight under braking, going wide in turn. Default setup of 99% of road cars and considered safer in general for average drivers. Oversteer = rear end of car comes around, usually resulting in a spin, spins remove all control from driver. Alert drivers in well handling cars can countersteer into the slide sometimes, this is (basically) drifting. Generally, only sports and racecars are setup this way.
A) they don't identify which bacteria is in the Yogurt. B) Lab testing has shown that yogurt is a terrible medium for supporting probiotics so most or all of them are dead by the time you eat it. C) Probiotics in foods don't typically establish themselves in your gut, it already is saturated with bacteria ideally suited to the environment. You would need frequent and high concentrations to disrupt that.
"To this aim, we conducted a prospective longitudinal interventional study in 21 healthy human volunteers not consuming probiotics"
Out of 21 they selected 15 for the study.
If you actually got any benefit from knowing your genetic conditions you were more pre disposition to then insurance companies would offer it to you for "free" to improve there screening process (or they could be purchasing that data from 23andme). But really knowing you are 20% more likely to get breast cancer are you going to start getting more breast exams, or just keep on living life?
Plus a couple months ago HN had a link to a news article where they sent there DNA off to multiple companies that do DNA testing and had mix results across the board. Companies like 23andme are only as good as the database of DNA they have. You will always get mix results because comparing DNA falls into NP where you can not improve the time it takes to compare DNA.
You will never get a full picture of you DNA from a service that only charges a few hundred bucks. It is like paying $10 a month and expecting a CDN at google scale.
I consider it worthwhile, even for me - "no news is good news".
I'm not even sure there is a one-size-fits-all answer to such things. That particular case seemed to be obviously one of greed, but most cases aren't as obvious. The most obvious (to me) would be some sort of price justification on new drugs, which takes into account r&d (to include a portion of failed drugs, but not advertising), manufacturing, and other actual costs with a certain percent of profits - with increases only rising with inflation or by going through the justification process yet again (to account for changes in manufacturing costs, for example). I don't see that happening soon, though.
DEF CON 25 - John Sotos - Genetic Diseases to Guide Digital Hacks of the Human Genome
idk the Chief Medical Officer of Intel considers genetic testing in general a pretty damn huge security risk and you completely neglect the whole bio weapons aspect.
1) Being arrested for real or framed crimes. In case you don't know, once someone has your sequence they can easily produce craploads of the regions checked by the police in bacteria and leave it at crime scenes.
2) Being denied certain jobs (eg GATTACA).
3) Being charged more for insurance, etc. In general you will start from a poorer negotiating position than someone who has your genetic data.
Of course, people think its great when it was used to solve a murder...
The paradoxical nature of genetic variation, especially deleterious variation, is that most of it is very rare. Even with today's massive genome sequencing efforts it's still common to find variants that don't exist in ANY previously sequenced genome. Looking only at just a few sites will rarely get you the whole story.
I do understand that the 'value' in DNA testing is largely the ability to follow the social graph of inheritance along with phenotypical metadata. This is pretty much at direct odds with privacy even for the strictest of 'anonymizing' setups considering that the greatest context-dependent 'utility' such as curing a truly unpleasant genetic disease or persecuting an ostracized person or group is due to identifying and understanding highly anomalous genes, phenotypes, or linkages of interest that come down to damning specificities.
There is one case I can think of that I _might_ be willing to participate in the "Human Genome Revolution", and at first thought it seems to me to be useful to both those that submit their samples and to humanity as a whole: a 'simple' counter of every allele across all DNA samples uploaded. However, I would appreciate feedback as to what are the gotchas here that will be abused, because I'm positive I have not thought such a system through at all and the Monkey's Paw [0] will grant my wish. :(
It would be a project that requires three components listed in increasing importance:
1. Cheap commodity DNA sequencer that can export its data in a free format. 2. A network protocol for uploading DNA alleles to a swarm of peers that archive and distribute the total counts. 3. A probablistic anonymizing data structure that combines something like a generalized crypto accumulator[1] and a zero-knowledge negotiation process between the network and a client for incrementing the common counters above a fairly large threshold such as maybe several thousand or million submissions. Perhaps uses a modification of something like the approximate counting algorithm[2] in a game setting.
[0] https://en.wikipedia.org/wiki/Monkey%27s_paw [1] https://en.wikipedia.org/wiki/Accumulator_(cryptography) [2] https://en.wikipedia.org/wiki/Approximate_counting_algorithm
If you have an interest in ancestry and family trees in can be useful. I've connected with a couple relatives I didn't know I had because of it. Trying to build out a family tree was a fun exercise for me and from a cost-time ratio I've gotten more out of it than I did movie pass.
Several people here seem to have been earlier adopters when they had more health related information that the FDA made them pull. They've been re-introducing some of them but for now 23andme is still a cheap way to get a useful sampling of your dna that you can run through services like https://promethease.com/.
If you get lucky (or unlucky depending on your perspective) it can make a difference in your life. It showed me that I had a slight increased risk of Celiac Disease so I had my allergy-immunologist do a blood test which came back positive. Endoscopy w/ biopsy also confirmed. I didn't have a lot of the stereotypical GI symptoms but after going Gluten free I can tell that I had other issues because of it. I doubt I would have been diagnosed correctly without my health report from them.
My wife also got a lot of value from hers. On the other hand our two adult sons got a couple mins entertainment from it and moved on.
That first line above sounds useful, right? But read on...
Then you find out that most people have a 0.3% chance of getting disease X (3 tenths of 1 percent chance, for those who wonder if I've made a typo). So, instead of a 0.3% chance, you have a 0.333% chance.
Big deal.
The ancestry part is interesting. But if I had it to do over again, I might not do it, because by doing so I made some of my family members uncomfortable by revealing some of their private information without their consent.
DEF CON 25 - John Sotos - Genetic Diseases to Guide Digital Hacks of the Human Genome
The Chief Medical Officer of Intel says genetic testing in general seems like a terrible security risk one should avoid incurring at all cost.
So that's a big fat no.