Having no statistical or genetic education, I can't completely follow your argument, but it sounds convincing. From what it sounds like, the article shows that there is genetic variation between ethnicities, but says nothing about how these variations, translated into behavioral variations, compare to cultural/circumstantial variations within a single group.
I'm not sure I follow your question. Which of my several statements do you have in mind?
Using the same markers, you can also pretty successfully predict the test subject's last name. This is interesting, perhaps useful, and it tells us something about the rate of non-paternity events (pretty low), but it shouldn't change anyone's worldview in a major way.
It is still not clear whether it is now considered not true that the variation within one ethnic group is larger than the variation between ethnic groups. This, I think, is true, and has not much to do with the fact that we can now reliably distinguish between ethnic groups based on genetic data only.
I was sloppy in describing of what was done, sorry. They generate genome wide variation matrices, diagonalize them and take the two largest eigenvalues, then plot the variation along the corresponding eigenvectors. These ARE the largest variations. The largest variations ARE ethnical. I think this is self-explanatory. Perhaps you have some other metric in mind, maybe you can formulate what it is, then we can talk numbers.
Have you seen a Nature study about genetic structure of Jews, published just this week? You probably did, but I thought of leaving here this link anyway, it continues our discussions here.
I've read it this weekend. Figure 2 is very impressive illustration of the use of PCA, isn't? I am less impressed with their structure analysis as various ad hoc assumptions go into it. But it tells the same story -- and this is the familiar story. I hoped for more, to tell you the truth. I hoped it would shed light on whether the Ashkenaz was populated from Northern Africa via the Mediterranian or from Persia, or in some other way.
Comments 21
Reply
Reply
Reply
Using the same markers, you can also pretty successfully predict the test subject's last name. This is interesting, perhaps useful, and it tells us something about the rate of non-paternity events (pretty low), but it shouldn't change anyone's worldview in a major way.
Reply
It is still not clear whether it is now considered not true that the variation within one ethnic group is larger than the variation between ethnic groups. This, I think, is true, and has not much to do with the fact that we can now reliably distinguish between ethnic groups based on genetic data only.
Reply
Reply
http://www.nature.com/nature/journal/v466/n7303/full/nature09103.html
The genome-wide structure of the Jewish people
Reply
Reply
Leave a comment