GENEALOGY-DNA-L Archives
Archiver > GENEALOGY-DNA > 2003-09 > 1062485700
From: "Cecelia Clancy" <>
Subject: [DNA] African Ancestry (was Cecelia's Mother's DNAPrint 2.0 Results - a big surprise)
Date: Tue, 2 Sep 2003 02:55:00 -0400
References: <20030902044244.13852.qmail@web41208.mail.yahoo.com>
Yes, revisions are laudible. When revisions are made, the company should
rename the product accordingly, such as DNAPrint 2.0 --> DNAPrint 2.01 --->
DNAPrint 2.01a ---> DNAPrint 2.2, and so on.
Another thing that bothers me about the way the results are reported is that
in all the results I have seen so far, DNAPrint ignores whatever the results
are for African ancestry - unless the person has majority African ancestry.
The sole exception I have seen for this is the results of Mark Shriver. He
is a DNAPrint person who considered himself to be "all White" but yet got
back a result of 20% African. No other specific percentages were given for
his IE, NA, or EA. (Today, on the DNAPrint website, his results have been
revised to 68% Indo-European, 16% African, 16% Native American. East Asian
is not even mentioned.)
They say they cannot visually plot four ancesties on a 2 dimensional paper.
But I have thought of a way to do it. My way won't lead to a specific point
(the little red dot that they call the MLE - most likely estimate), but it
will show a range. I'll explain my proposed method later. And they should
also make the raw numerical results available for all four possiblities.
This is particularly irksome to me because I am specifically trying to find
out whether my maternal grandfather's recent (past few hundred years)
ancestry included African, Native American, or both, (or now, if my Mom's
100% IE is indeed correct, "or neither."). So, I would really like to see
how much or how little Mom's confidence interval leans towards the African
pole. DNAPrint at present, is presenting only part of the picture -- in
other words, reporting only part of what the data shows.
I am also wondering whether some people at DNAPrint "tweaked" the algorithm
to lessen the strenght of the African signals.
The reason is that when DNAPrint first announced its Ancestry By DNA test, I
found out very quickly and was one of the first people to sign up. At the
same time, I also read all of the articles that were on the Ancestry By DNA
website at the time.
Under the "Case Studies" section, there was information about Tony Fudakis's
results, as well as the results of his wife and of one or more of his
children.
I distinctly recall that back in those very early days, Tony's result was
reported to include 6% African. Tony also was reported to have majority
Indo-European and significant Native American ancestry. He had no East Asian
at the time. His wife had a mixture of Native American, Indo-European, as
well as some African. She showed no EA.
They explained his wife's African result by reporting (correctly) that there
is African genes in the Mexican gene pool. (They also included - and still
include - an article on this in their Articles section - it is called
"Mexico's Missing Blacks.")
They admitted being baffled by Tony's 6% African, but said words to the
effect, "it probably has to do with his Greek ancestry" and made no further
comment. They did not add things like - the sickle cell trait is more common
in some villages in Greece than it is among African Americans.
They did not speculate how that sickle cell gene got to Greece in the first
place. Like, perhaps in ancient times, people from south of the Sahara might
have migrated up the Nile River to Alexandria (today Cairo, approximately),
and from that town that was very important to ancient Greek culture, they
might have migrated to Greece and then over some generations became absorbed
into the general Greek population. In other words, they did not explain how
African might logically have something to do with Tony's Greek ancestry. I
got a sense that Tony was ashamed of it rather than amazed by it.
The triangle plots for all the members of the Fudakis family had at the
three points - African, Indo-European, and Native American.
Then weeks later, when I visited the website again to reread the Case Study
on Tony and his family, the 6% African for Tony was gone. In fact, *all* his
African was gone. So was all of Mrs. Fudakis's African gone. And the
triangle plots now had (and still have) at the points - Indo-European,
Native American, and East Asian. This in spite the fact that both Mr. and
Mrs. Fudakis have a 0% charted for the EA parameter.
Worse, the text under (or over) the graphs for both husband and wife said
"there was no detectable East Asian or African Ancestry." But yet, they
know, and are now seemingly hiding the fact that a previous test (same
publicly-stated version thereof), that both showed significant African
ancestry.
So, my best guess here is that they changed the algorithm to reduce the
strenght of the African signals - to attenuate them. But why? Was it because
they thought for purely scientific and mathematical reasons that the
original algorithm was faulty? Was it because Mr. or Mrs. Fudakis was
"ashamed" of being of African ancestry??? Was it because relatives of Mr.
and/or Mrs. Fudakis's objected to being publicly "outted" as being
"Negroes?" (Remember, to many people in America even today - "One drop of
Negro blood makes you all Negro.") Or, did kids in the neighborhood or in
school start calling their children the n-word???
So, why did Mr. and Mrs. Fudakis's African result disappear? Was it for
scientific reasons or for social reasons?
And, I am wondering if they got angry letters from racist Whites who took
the test and their results included some African? If so, this would be
"motivativation" to attenuate the African signals in order to give the
customer what he or she wants????
I hope not. I hope the disappearances are for valid scientific reasons. I
suppose twenty years from now when the patent runs out - we can run the data
through our own algorithm and see what we get.
And again, if they are indeed changing the algorithm (as somebody at
DNAPrint told me explicity over the phone that they did so for my second
test), they should let these revisions be known by making minor changes in
the name of the test, such as changing the original DNAPrint 2.0 to
something like DNAPrint 2.01.
Regards,
Cecelia
This thread: