[ExI] 23andme again

Alejandro Dubrovsky alito at organicrobot.com
Mon Jul 1 13:30:58 UTC 2013


On 01/07/13 22:47, Mirco Romanato wrote:
> Il 01/07/2013 05:48, spike ha scritto:
> 
>> Daaaaaaaam!   It's happened again!  And this time it was an accident.
> 
>> 23andMe gives you a list of DNA relatives sorted by how close they are
>> genetically.  It also gives you a list sorted by what they call "enrichment"
>> which divides thru by how common is the name.
> 
>> So last night I drop the guy a note along the lines of Hi cousin, I see you
>> are descended from William Yakity Yak and Mary Bla Bla.  This morning I get
>> a note back saying he doesn't have either of those in his family history
>> database which goes back four generations.
> 
>> OK so do I post back and say: Ooops sorry my mistake, or do I post back and
>> say: Cousin, you have a major error in your genealogy, or do I post back and
>> say:  Pal, one of your fairly recent ancestors was apparently adopted?
> 
> I suggest the:
> 
> Maybe 23andMe is wrong or there is an error in my or your genealogy.
> What could be?
> Do you want try to exchange data?
> 

I think the possibility that 23andme is "in error" should be considered
seriously, not just as a method of releasing information on unsuspecting
relatives. I put the quotes around "in error" because I mean not just
lab-handling errors, but historical anomalies that make genetic tests
indicate closer kinship distance than there is in reality ***.

For example, it is well known (TM) that most of the relationships shown
on 23andme for Ashkenazis are rubbish due to the deep inbreeding in that
group. eg from their blog
http://blog.23andme.com/news/announcements/how-many-relatives-do-you-have/,
they say "we estimate that any two randomly chosen individuals who
identify as Ashkenazi are on average the genomic equivalent of 4th-5th
cousins, because they share many recent common ancestor". 23andme will
list scores of 2nd cousins which you are pretty sure are nowhere near
that close.

My guess is that there is likely to be other groups that are similarly
affected. Any geographically isolated community, or any small,
long-lasting religious group should show similar effects.

*** reality here is probably the wrong term, but I assume that, for most
people, someone that, due to historical reasons, shares the same amount
of genetic material as a second cousin doesn't have the same emotional
or gossip-al power as someone you share a great-grandmother with.



More information about the extropy-chat mailing list