Posted by Kate Phizackerley on Sunday, February 28, 2010

I've not yet waded through the full JAMA paper, let alone the genetics papers I need to read so that I can fully understand the detail, but I am still at the stage where what I am reading is increasing my concern.  For instance the JAMA paper indicates that all of the tested mummies had either allele 26 or allele 27 at the D2S1338 locus. The paper Population Data on the STR Loci D2S1338 and D19S433 (Bruce Budowle et al, 2001) reports distributions for this locus.  If you refer to Table  1 you'll notice that characteristic 27 was not observed in the African American, Caucasian, Hispanic or Filipino populations sampled and was rare in the Chamorro population.  Characteristic 26 was the second rarest.

Yet in the JAMA paper every one of the tested mummies has one of these characteristics:

Mummy                          D2S1338
Thuya (LV46)                   19,26
Yuya  (KV46)                   22,27
KV35EL                         22,26
Amenhotep III (KV35)           16,27
KV55                           16,26
KV35YL                         16,26
Tutankhamun (KV62)             16,26
KV21A                          N/K,26
KV21B                          17,26
Foetus 1 (KV62)                16,N/K
Foetus 2 (KV62)                N/K,26

The DNA sampling is incomplete where N/K is shown.

Now it would seem at least possible from this, perhaps even probable, that the allele distribution of the royal family has narrowed so much that all members possess either allele 26 or 27.  Even the other distribution of the other characteristic is narrow.  It is my contention that it is reasonable to use the normal general population distribution probabilities (unless it is subsequently shown that the genetic variance of the entire nobility is narrowed) when showing consanguinity, but when considering particular relationships within the population such as father son etc, then probabilties based on the allele distribution of this poulation should be used.  I'm intending to dump the whole table into Excel and see what I can manage in terms of probabilities.

It's also interesting that, with the exception of the KV21A mummy, foetus one must have at least one parent other than those on this list - ie the source of the 17 variant.  (It could be mummy KV21A because one allele for her is unknown.)  That's just based on one allele, of course.  (Impossibilties can be established by one allele - if we ignore the chance of genetic mutation.)

1 comments:

Anonymous said...

I`m sure it would be very interesting if you can put the distribution of alleles into the context of this population.

Something else concerns me which has been mentioned in Pling`s comment on the DNA-testing process. It is the fact that only 8 markers are shown for the testing of the nuclear DNA. In the JAMA paper they talk of "all 8 markers yielding results", indicating that these were the only ones tested.
As pointed out by Pling this number is rather low and usually up to 16 markers can be tested (which they did with the Y-chromosomes).
In one of the short videos on Youtube ( the one is called "Maternal DNA Match") the speaker briefly mentions that the 8 markers "they could read" provided the crucial matches between the DNA of Tut,KV55 and the YL.This sounds as if in fact more that 8 markers have been tested but some of them unsuccessfully.

Cosidering that one cannot know what kind of information additional markers would have provided and that modern courts usually require at least 10 matching markers as a "proof" for parenthood I wonder how reliable the presented results really are.

Hawass would do us a great favour if he allowed another completely independent institiution to repeat the testing.

Search

Admin Control Panel