Stories
Slash Boxes
Comments

SoylentNews is people

posted by mrpg on Friday August 18 2017, @12:00PM   Printer-friendly
from the color-me...-anything dept.

Over at StatNews is a story on a recent trend where low cost commercial DNA testing is resulting in a number of White Nationalists taking genetic tests, and sometimes they don't like the results that come back.

The article looks at research on how they respond to the sometimes unexpected results:

[...] In a new study, sociologists Aaron Panofsky and Joan Donovan examined years' worth of posts on Stormfront to see how members dealt with the news.

[...] About a third of the people posting their results were pleased with what they found. "Pretty damn pure blood," said a user with the username Sloth. But the majority didn't find themselves in that situation. Instead, the community often helped them reject the test, or argue with its results.

Some rejected the tests entirely, saying that an individual's knowledge about his or her own genealogy is better than whatever a genetic test can reveal. [...] Others, he said, responded to unwanted genetic results by saying that those kinds of tests don't matter if you are truly committed to being a white nationalist. Yet others tried to discredit the genetic tests as a Jewish conspiracy "that is trying to confuse true white Americans about their ancestry," Panofsky said.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 2, Informative) by Anonymous Coward on Friday August 18 2017, @02:19PM

    by Anonymous Coward on Friday August 18 2017, @02:19PM (#555883)

    Yes, but the error rates (usually ~10^-8 per bp/division each) are too low for it to peak so early. That is why in the original paper they say the probability of accumulating n mutations occurring at rate p by* time t is (I am setting p_1 = p_2 =... = p_n here for simplicity):

    A = (p*t)^n

    This is instead of using (more computationally intensive) geometric distribution:

    B = (1 - (1 - p)^t)^n

    If you check, you'll see the approximation A ~ B works only for low mutation rates (p [much less than]1):

    This result will be valid for large values of t (of the order of a human lifetime) provided that p1t, p2t, . . . , prt are all sufficiently small (as could be assumed in an application of this theory to human cancer).

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2007940/ [nih.gov]

    They never consider that the assumption that p [much less than] 1 is incorrect, instead coming up with all sorts of other explanations:

    In later analyses, however, it was noted8 that at higher ages the age-specific rates showed substantial departures below the rates predicted by log-log behaviour. Doll and Peto9 reported that lung cancer risk among men aged 80–84 appeared to be half that among men aged 75–79. They offered four possible explanations for this finding. These were under-diagnosis, selective survival, unreported cohort differences of smoking patterns in early life, and the possibility that the biology of extreme old age reduces the risk of carcinoma.

    https://academic.oup.com/ije/article/33/6/1182/866607/Commentary-Fifty-years-of-the-multistage-model [oup.com]

    *Note that we are working with the cdf here, to compare to age specific incidence you need to get the pdf (first derivative or first finite difference of the cdf, or in the case of the approximation multiply by p*dt for the probability of getting the last mutation in interval dt).

    Starting Score:    0  points
    Moderation   +2  
       Informative=2, Total=2
    Extra 'Informative' Modifier   0  

    Total Score:   2