The combination of access to genetic data and computer programming must be an irresistible temptation to racists. Someone tried to distribute a code fragment that would allow a program to look up gene data on 23andMe and use it to limit who would be allowed to use the program.
Dubbed Genetic Access Control, the program—which was posted to GitHub on Monday—would act as a login for sites and scan the genetic information of 23andMe users who make their data available, much like how websites currently request access to your Facebook profile prior to entry. The coder in question cited a few “possible uses” for Genetic Access Control, ranging from “Groups defined by ethnic background, e.g. Black Panthers or NAACP members,” to “Safer online dating sites that only partner people with a low likelihood of offspring with two recessive genes for congenital diseases.”
Or Stormfront could use it to make sure no one with one drop of black blood could register for their forums!
Fortunately, 23andMe has restrictions that don’t permit its use in such a way, and they’ve already taken steps to block “Genetic Access Control” from using their databases in any way.
I like how it’s just casually thrown in there that the NAACP restricts membershiop based on race, as if that were remotely true. Anyone of any race can join the NAACP and no one is disallowed to hold any position within the organization based on race.
There goes Craig Cobb’s stormfront membership.
Seriously, a perverse part of me hopes they’ll implement it. There’s probably a lot of racist out there who don’t realise that great-grandma was black.
I’m just waiting for the INVALID-signs to start flashing everywhere.
Once such a system is in place and protecting sites of interest or value, suddenly there’s a financial incentive for people to break the system. Just spend a few bucks to purchase a small amount of someone else’s bodily fluids, and use those to register yourself on the appropriate website. Or just get someone else to sort out the messy bit and simply provide you with login credentials. It’ll be like buying twitter followers.
What exactly is 23andMe? I mean, I understand that it’s some sort of genetic database, but why does it exist? I’ve looked at their website, and the only useful things they seem to offer is telling you whether or you might pass on a genetic disease to your children, but who pays £125 to find out whether or not they might pass on a inheritable disease? If you were at risk of doing so, you’d know from your medical records, surely?
@ Emerson White #1
And Black Panthers =/= All Black people. The reasons given for creating this program are pathetically transparent.
Oh, good. So not racism. Just eugenics. That’s totally okay, then.
@Thumper #5:
23andMe is a “for entertainment/information only” genetic screening service. You send them a sample, (and some money), they screen it for a bunch of genetic markers and send you back a report that gives a low-grade guess as to where your ancestors came from and a next to useless listing of possible predispositions for illnesses.
@ Rick Pikul
I suspected that their gumpf about detecting genetic illnesses was overstated. So basically, they do what any family-tree-tracing company does through birth certificates etc.; but with some fancy technology at twice the price.
Genetic Access Control to make online dating “safer”? What a clear indicator that someone has no fucking idea the ways in which online dating is currently, actually, unsafe.
And “only partner people with a low likelihood of offspring with two recessive genes for congenital diseases?” One, good luck if you have a large enough pool of “congenital diseases” that you are actually looking at. Two, it is rather invasive to ONLY allow partners reaching a certain arbitrary low likelihood threshold. Three, this is online dating. Just dating. Not even that, just planning to talk and potentially set up a first date. Not plans to get married, let alone plans to have a child together.
And then there is the eugenics and racism behind all of the proposed uses of this. Who the fuck thinks this would be a good idea?
They way I first read it, I thought they were saying that the Black Panthers and NAACP are an ethnicity. I think what they mean is that such groups would want to screen people on the basis of their ethnicity to ensure that only black people could login. As previously noted, the NAACP wouldn’t do this. As for the Black Panthers, they’ve been defunct for over 30 years. But the important thing is that we single out black people as the discriminatory ones.
I would be seriously surprised if they cared more about any possible use for their screening service was put to or who would use it than they care about all the money they hope to make out of it.
their dreams are more likely not dreams of race purity as much as they are of a Mark Zuckerberg fortune.
uncle frogy
Better scan for racism or bigotry, if scanning for anything at all.
@Thumper at 5 and other comments:
23andMe is specifically prohibited from promoting their service as a way of identifying and assessing disease risk, since there is no demonstration that their sequencing is either sensitive or specific enough for tracing known disease-causing mutations. Nor do they provide any of the other information and expert assessment needed to actually interpret their results in a clinically meaningful way. The team over at Science Based Medicine have written about this: https://www.sciencebasedmedicine.org/the_fda_and_personalized_genetic_testing/ .
As you say, personal and/or family medical history is usually a reason for genetic testing. e.g. I know that I carry a predisposing allele for hypertrophic cardiomyopathy, because there is a family history of the disease indicating that potentially-at-risk relatives be tested. There are some situations where genetic testing can be useful on a population-wide scale, absent immediate family history (e.g. Tay-Sachs carrier status), but that’s something to be done by genetic counselors and MDs with the relevant specialty expertise and not by a group doing unverified sequencing by internet mail-order.
Re. This particular bit of bigoted evilness:
If the coder concerned is based in the United States, it seems likely that they just violated the Genetic Information Nondiscrimation Act: https://en.wikipedia.org/wiki/Genetic_Information_Nondiscrimination_Act and possibly a few other laws. If so, may they and anyone who tried to use their program find themselves on the receiving end of an appropriate lawsuit.
The code is still at GIThub (do I need to add a link?) and I downloaded a ZIP-file (SHA-256 checksum: 4764415d3ca193af3a4cb2fb4c63dd03f591f6f71862b5e817980d287390ba15) of it to take a look at it.
It is a simple Python script that uses the API available from 23andme to get some informations of a specified user of 23andme. These informations are very specific by the way and the choices are very telling. Also very telling is the programmer’s pseudonym: “offensive-computing”. He is the single contributor to this repository, at least for now.
The script starts with some hard-coded (meaning you can only change them by changing the code) variables. Most are just basic things, like the address of 23andme’s website, the interesting variables start with the setting of some aptly named thresholds.
allowed_population_threshold = 0.51 # minimum allowed match %ancestry_speculation_threshold = 0.75 # standard ancestry speculation(The hash-sign starts a comment and that the comment says the same as the variable is a thing that the people over at TDWTF make fun of, I will spare it tis time) I did not check what the numbers mean or if the are meaningful at all in the first place.
The next one, well…
ancestry_allowed_populations = [ 'French & German', 'British & Irish', 'Finnish','Scandinavian', 'Northern European', 'Eastern European', 'Balkan', 'Iberian',
'Italian', 'Sardinian', 'Southern European' ]
And if you think that is, how shall I put it politely, highly questionable, the comment following this line is:
# note: does not include "Ashkenazi" or "European"In the short list of “Possible Uses” in the file
README.mdthe second point is:(The urge to wash my hands is imminent now)
The whole thing tries very hard to sound reasonable and neutral with it’s pros and cons in the readme-file but the code itself does not lie:
The script works by downloading the users profile, sorts the information regarding ancestry out if the threshold in
ancestry_speculation_threshold(75%) is met, checks the content against the list inancestry_allowed_populationsand if the total in percentage is above or equal to the threshold defined inallowed_population_threshold(51%).It does some very basic checks for errors, too, although very basic checks of all but a few possible errors.
Yepp, that’s all it does: a check for “whiteness” with a not very hidden hint that it can check for “jewishness” as well.
The script has no explicit license at all, which means nowadays that the programmer holds the full copyright and you cannot use it without the programmers consent. The script is based on the example script of the API of 23andme which also has no explicit license. I hope he has the (preferably written) consent of 23andme to use it.
GATTACA
Oh please, YES.
…all of the sudden, half of the racists discovered their lineages to be a lot more complex than they expected…
Users would actually allow this?! That’s the part that surprises me.
@ Area Man #11
Ah, I read it as “This program can help to deny entry to groups like the Black Panthers”, when of course what it would really do is deny access to black people. It’s entirely possible I read it wrong, though.
@michaelbusch 14
Well, the genotypes they collect are highly likely to be accurate. But that still leaves two problems
1) They are only looking at known variations, so any mutation unique to you, or unique to your family, they can’t see.
2) The state of biology is such that even if they see that you have a rare allele, we do not know enough biology to say “Oh, that variation predisposes you to this health consequence”. Lots of times people publish papers saying “This SNP correlates with that health condition”, but most of the time, other studies of different people will not replicate the finding, that is, people are mostly publishing coincidences. It’s bound to happen occasionally, when people use crappy statistics, it happens a lot more often.
I would guess that their ancestry claims are pretty well backed by solid data. But the health claims, those are backed by pretty much nothing, and they weren’t even trying to have a biologist with a brain look over the claims they were sending out.