The consumer genetic testing company's official blog, covering its research findings, product updates, regulatory news, and security notices.
HN commenters link it for specific findings like the cilantro-aversion gene, and in discussions of the account breach where credential reuse and company responsibility were debated.
you say “obviously” there is no genetic criterion for being Latino... there’s nothing obvious about that at all. you say here yourself they aren’t a random sample of the population - then what are they? and whatever it is that ties these individuals together, I’d wager there’s a genetic history overlapping in those tying elements. Or else you’d find much more ancestral diversity within “Latinos” (or “Asians”, or “Blacks” or “Italians”. there’s a reason these partially-cultural-mostly-ancestral groups _tend_ to have similar ancestry. tend.)
just one example of the genetic angle of Latino-ness: https://blog.23andme.com/ancestry/latino-ancestry/
if you go far enough back, we’re all african, and if you go far enough forward, we’re all exactly X% everything after enough gene-mingling. but we’re not at that X% mix and we’re also not all african. we’re in an in-between stage where the cultural groups that are typically called “races” in this in-between moment on average tend to share ancestry and as a result, on average tend to be genetically similar.
why are you making this sound like it’s a wild claim?
I think of it as similar to taxonomy of species - there’s no line during evolution where something stops being one species and starts being another. just like there’s no hard line between races. but a sparrow is clearly not a giraffe. that’s a useful comment about their genetic differences that may or may not give you information related to the correlations of those…
Thank you for calling this what it is - a hack - despite 23andme's strenuous efforts to paint this as the fault of millions of users (see https://blog.23andme.com/articles/addressing-data-security-concerns) rather than owning the vast technical or management failure that allowed this to continue undetected for months.
Without the risk of a giant fine or, say, jail time, many tech giants can and do get away with managing their data security badly.
That's right. It's happened before, and will continue to happen as long as there are no consequences.
Note that 23andMe is not the first online genealogy service to get hacked:
- In 2017, MyHeritage had 92 million accounts hacked https://www.hackread.com/dna-testing-website-myheritage-hacked-user-accounts-stolen/.
- In 2020, MyHeritage users were targeted in a separate phishing scheme. https://blog.myheritage.com/2020/07/security-alert-malicious-phishing-attempt-detected-possibly-connected-to-gedmatch-breach/
- GEDmatch admitted “all user permissions were reset” in a 2020 attack. https://www.buzzfeednews.com/article/peteraldhous/hackers-gedmatch-dna-privacy
- Ancestry and Ancestry affiliated companies have had multiple security breaches over the past 10 years (https://www.hackread.com/software-firm-leaks-ancestry-com-user-data/)
- Ancestry has also destroyed people's archives when it decided it was no longer profitable or important enough to keep them.…
The only thing 23&Me has done really well in stat gen is IBD (which IIRC you worked on).
I was a team from Google that evaluated 23&Me's data and technology many years ago for a business deal. We already talked about this with Anne and she confirmed what I said above. It might have been before you were hired- but I'm pretty sure we talked about building a variant store and a transpose service? I stand by my statements (note: I work at a competitor of GSK and I know all about these deals). It's not correct that being better at picking targets doesn't shorten the timeline, either- at least in the opinion of the scientists at my company.
Additive models don't really account for most of the heritability of complex phenotypes. They're what have worked best and been published so far. Complex phenotypes are nonlinear because the generative processes in biology have feedback, homeostasis, enormous numbers of individual elements... etc...
By the way, does 23&Me still put tongue-twisting as a heritable trait on its reports? https://blog.23andme.com/articles/tackling-tongue-curling-the-challenge-of-23andwe It's not.
I want to look into something that is truly cutting edge and can cause disruption.
John listed many good projects. Based on ^^, other good places to look to help generate ideas:
Current notable movers:
* http://rockhealth.com (also check out their Events page in case you may able to get cheap access to conferences as a student)
* http://blog.23andme.com
And places to see what people are talking about/building:
* http://www.quora.com/Health-Informatics and all sorts of related topics
* https://angel.co/health-care-information-technology and all sorts of related topics
* http://mobihealthnews.com
* http://www.imedicalapps.com
And in general I'd recommend following any company/person you learn about from the above ^^ on http://crunchbase.com
Also feel free to shoot me an email tyler@thetylerhayes.com. I'm in the process of starting up my own health tech-related project (http://blog.stayinyourprime.com) so I've been doing a lot of research. Happy to share what I've learned and chat more about what you're doing.
Conversely I supplement my PCP with a cash pay MD (in a hospital setting with a background from a reputable school) who believes in "wellness" which he views as the flaw in western medicine. Instead of going from sick to ok, he believes we should go from ok to wellness. And he sells mega Vitamin D doses and maybe I'm a sucker but I buy them.
Since its cash pay and I absolutely love testing/scanning myself randomly (any blood/dexa/cortisol/ekg/heart test I can find and afford) we usually spend 20 minutes twice a year going through the science of whatever I'm testing at the time and (typically) shooting down the necessity of whatever crazy thing I want to try next.
But low impact low benefit things like vitamins and a few no-cost hobby things are super under-rated. For example theres a handful of entries on this chart that are low cost with low behavioral change required medium impact that were unexpected for me: https://blog.23andme.com/health-traits/what-patients-say-works-for-depression/
I wonder why its unethical to test such things on human beings. I'm assuming it's unethical, because there rarely seems to be any damn work on this using humans.
Those rats ran themselves to death, will we humans?
We share about 92% of the genes[0] (on the first google attempt), what changes does that giant 8% contain?
Just pick up people like me, and make us run and give us food.
I doubt we will object if it means there's a possibility that we might help ourselves.
Are they afraid they'll make it worse? Sure that's a possibility, but that's the price to pay. I'm willing to pay that price if it means people after me, who have this wire crossed in their brains will be served better healthcare in the future.
Some experiments might not work, but we've just discovered another way to _not_ make the lightbulb, but the good news is that the lightbulb _will_ be made if we try n times.
[0] https://blog.23andme.com/23andme-and-you/genetics-101/genetic-similarities-of-mice-and-men/