Log in

No account? Create an account
19 December 2012 @ 04:38 pm
The Death of Ramesses III and Revisiting Tutankhamun's Family  
A couple of days ago BMJ published a paper about the death of Ramesses III (Revisiting the harem conspiracy and death of Ramesses III: anthropological, forensic, radiological, and genetic study. Zahi Hawass, Somaia Ismail, Ashraf Selim, Sahar N Saleem, Dina Fathalla, Sally Wasef, Ahmed Z Gad, Rama Saad, Suzan Fares, Hany Amer, Paul Gostner, Yehia Z Gad, Carsten M Pusch, Albert R Zink. BMJ 2012;345:e8268). This uses techniques similar to the 2010 paper which gave details about the health at death and the familial relationships between several of the late 18th Dynasty Pharaohs (Ancestry and Pathology in King Tutankhamun’s Family. Zahi Hawass, Yehia Z. Gad, Somaia Ismail, Rabab Khairat, Dina Fathalla, Naglaa Hasan, Amal Ahmed, Hisham Elleithy, Markus Ball, Fawzi Gaballah, Sally Wasef, Mohamed Fateen, Hany Amer, Paul Gostner, Ashraf Selim, Albert Zink, Carsten M. Pusch. JAMA. 10;303(7):638-647). In this case they concentrated primarily on one mummy, that of Ramesses III, and investigated his cause of death. They also looked at whether or not another mummy could be a son of Ramesses III.

Ramesses III was a 20th Dynasty Egyptian Pharaoh, ruling for around 30 years in the 12th Century BC. It's known from ancient Egyptian texts that there was a conspiracy against Ramesses III, a plot to assassinate him and for his son Pentawere to become Pharaoh. The documents record the trials and punishments for the conspirators, and it is clear that the coup failed. It isn't clear from the text whether or not Ramesses III was actually killed. The Egyptians believed that writing things down was magic and that it made a statement true. They are often rather ambiguous about things like the death of a Pharaoh, particularly a violent one, so the lack of explicit mention of Ramesses III's death doesn't really give us much information about whether he did die or not. One of the reasons this paper is interesting is because we already know so much about this piece of politics from 3000 years ago, and there's still more we can learn with new technology.

The mummy of Ramesses III had previously been inspected for signs of a violent death but none had been seen with the techniques available at the time. This paper uses CT scans of the mummy and shows that underneath the bandaging around the throat there is a large cut (7cm across, gaping 3cm wide and cuts through all the soft tissue from the front right through to the vertebrae at the back). If it was done to him while he was living it killed him. Not only is the neck heavily bandaged to hold it together but also there is an Eye of Horus amulet in the wound (which would be a healing charm). The authors conclude from the evidence that it was almost certainly the cause of death, rather than being inflicted during the embalming procedure. And they speculate that the amulet is there to enable the king to heal as he enters the afterlife.

They also looked at another mummy - which hasn't been mummified "properly" and was wrapped in a goat skin (which is a ritually impure way to do things). Yet he was still preserved (and even moved to the cache of royal mummies in Deir el Bahri, so treated with some respect). They were interested in seeing if this was Pentawere, was this a part of his punishment for conspiring to kill the Pharaoh. They did use CT scans to look at the body, it is that of a young man age 18-20 and there are signs of a death by suffocation or perhaps strangulation although this is not clear because of the poor state of the body. It's known from the trial records that Pentawere was found guilty and condemned to kill himself, which doesn't quite match with suffocation/strangulation but perhaps he hung himself? But they say that none of this is really clear. They also did genetic analysis on the two mummies, and showed that they are potentially father & son. Because they don't have the mummy of Pentawere's mother (Tiy) available they couldn't show anything for sure - but it seems likely from the data that this is a son of Ramesses III and from the way the body was treated it only really makes sense for it to be Pentawere (based on what we know).

I'm particularly interested in the genetic data because it can be used to answer one of my criticisms of the 2010 Tutankhamun paper (I wrote about the paper at length here). The methods in the 2012 paper state that the analysis was done the same way as in the 2010 paper, and I actually wonder if it was done at the same time (but that's pure speculation on my part).

I had a few criticisms of the way the data was reported in that earlier paper - there are no controls shown for the autosomal markers, the total number of usable samples for each mummy is not shown (nor the places they were taken from) and the number of samples giving the same data point is not shown (ie are 16 out of 30 in agreement, or 29 out of 30? both count as "majority" but one is more convincing than the other). Unfortunately most of these criticisms still hold true for the 2012 paper - one that doesn't quite is that they do list the places on the mummy where the samples were taken, although they still don't indicate which and how many samples actually yielded usable DNA.

The lack of controls in this 2012 paper for the autosomal data is particularly poor, because they publish some controls for the Y chromosome data (the control from the kit they use for the markers, plus 3 of the team members). It would be good if they also used another mummy for a control, but it's still better than the data presented for the autosomal markers. However, taking the two papers together you can use them as controls for each other (to some extent, it'd be nice to have data from the same run of experiments but this is better than nothing).

A quick run-down of how short tandem repeat analysis works (copied from my previous post on the subject). There are sections of DNA that are repeats of short sequences. It doesn't appear to do anything, and as a result these regions mutate more than functional sections of the genome, so can be used to trace familial relationships (unless of course you get a mutation right between two generations of the family you're looking at, but that is still a rare occurrence). For any given marker what they do is look at how many repeats the sample has - there will be 2 numbers of repeats, as chromosomes come in pairs. Brief explanation - the DNA in a person is organised into 46 chromosomes, 22 pairs of non-sex chromosomes normally known by numbers and 2 sex chromosomes (XX for women, XY for men). You inherit one of each pair of chromosomes from your father and one from your mother, and this is randomly organised. So if your mother's chromosomes are labelled A & B, and your Dad's are C & D, you could inherit chromosome 1 as A (from your mother) & C (from your father), chromosome 2 as A&D, chromosome 3 as B&D, chromosome 4 as A&D and so on; a full sibling of yours would get a different random selection, ie 1 B&D, 2 A&D, 3 B&C, 4 A&C. And importantly you MUST get one in each pair from each parent, you'll never have 1 A&B for instance.

So with the STR analysis you get a bunch of numbers for repeats of the short section for each marker, and you use markers spread across the chromosomes so that they are independently inherited. To look at one of the markers for Tutankhamun's family (a marker called FGA), Tutankhamun has 23 repeats on both chromosomes. Amenhotep III has 23/31, KV35EL (Tiye) has 20/26. So Tutankhamun cannot be the child of those two people - where would he get the second 23 from? KV55 has 20/23 - so you see he can be son of Amenhotep III/Tiye, and KV35YL has 20/23 as well (so she can be a daughter). And Tutankhamun can be the son of KV55/KV35YL because he can get one 23 from each of them.

Basically you do this sort of comparison across all the markers you look at - 8 in this case. And calculate probabilities for each trio (mother/father/child), they say 99.73% is the accepted standard for "practical proving" the relationship.

But this does rely on the population you're looking at not being so inbred that some markers are more similar within the population than you'd expect. You could possibly mis-identify a familial relationship - and that's why you should have negative controls from a non-related member of the same population. In the case of these papers the best control of all would be a selection of mummies from the elite population of the same time period as the test subjects, but not part of their family. But the rarity of finding named & dated mummies means that settling for an unrelated mummy from the elite population of a different era would likely be the best one could do.

And it does irritate me a bit that they didn't use a sample from the already published 18th Dynasty mummies as a negative control for these 20th Dynasty ones. Ramesses III ruled about 150 years after Tutankhamun died, and there's a couple of breaks in the line of succession during this time that means he's not directly related to Tutankhamun. But clearly still from a family that is part of the elite population. So I did the comparisons myself:

 D13S317D7S820D2S1338 D21S11D16S539 D18S51CSF1POFGA
Ramesses III vs ThuyaMatchMismatchMismatchMatchMatchMatchMatchMatch
Ramesses III vs YuyaMismatchMatchMismatchMismatchMismatchMatchMismatchMismatch
Ramesses III vs KV35ELMatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Ramesses III vs Amenhotep IIIMismatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Ramesses III vs KV55MatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Ramesses III vs KV35YLMatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Ramesses III vs TutankhamunMatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Ramesses III vs KV21AMismatch       
Ramesses III vs KV21B  Mismatch Match   
Ramesses III vs Fetus 1MatchMismatch      
Ramesses III vs Fetus 2 Match MatchMatchMismatch  
Unknown Man E vs ThuyaMatchMatchMatchMatchMismatchMismatchMatchMatch
Unknown Man E vs YuyaMatchMatchMismatchMismatchMismatchMatchMismatchMismatch
Unknown Man E vs KV35ELMismatchMismatchMismatchMismatchMismatchMismatchMismatchMatch
Unknown Man E vs Amenhotep IIIMismatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Unknown Man E vs KV55MismatchMismatchMismatchMismatchMismatchMismatchMismatchMismatch
Unknown Man E vs KV35YLMismatchMatchMismatchMismatchMatchMismatchMismatchMismatch
Unknown Man E vs TutankhamunMismatchMismatchMismatchMismatchMatchMismatchMismatchMismatch
Unknown Man E vs KV21AMismatch       
Unknown Man E vs KV21B  Mismatch Mismatch   
Unknown Man E vs Fetus 1MismatchMatch      
Unknown Man E vs Fetus 2 Match MatchMatchMismatch  

I compared Ramesses III and "Unknown Man E"'s data to each of the other mummies in turn, and colour coded the mismatches. If a given mummy didn't have data for both chromosomes then I skipped that marker because you can't tell. In a few cases I wasn't sure what counted as a mismatch - this isn't quite my field, and so I'm not sure exactly what 29.2 repeats means and if it's significantly distinct from 29 repeats. So those mismatches aren't coloured red.

You can easily see there's no way you could identify any of the pairs of mummies as in a parent/child relationship. So that's nice to see, and makes the data they do show in each paper a bit more convincing. It also makes the statement that they make in the 2010 paper, that Tutankhamun is possibly the father of the two fetuses buried in his tomb a bit more interesting - even with the patchy data available they don't match the controls.

If you're looking at that and thinking it's awfully dull, you're right. This doesn't tell us anything interesting, but I think it's important to know that the control data isn't interesting. And I do wish they'd actually published the controls with the data.
Fortune depends on the tone of your voicebopeepsheep on December 19th, 2012 08:09 pm (UTC)
That's actually really interesting, and probably the most understandable explanation I've heard in years. :)
Margaretpling on December 19th, 2012 08:56 pm (UTC)
Thank you! :)
(Anonymous) on December 31st, 2012 03:07 am (UTC)
Population Affinity

Thanks for the fine explanation; you make it lok almost too easy

I do have a question though: I ran the autosomal STR values in a popaffliliator and got for any other random no I chose in addition to the 8 STrs in the study, a reading of 'subsaharan African' at more than 90%. WHat I do with that info? I believe same result was gotten for the Amarna pharoahs some time ago.

Thank again
Margaretpling on December 31st, 2012 05:33 pm (UTC)
Re: Population Affinity
Within a population people are more related to each other than they are to other populations (i.e. any two random native Europeans are more closely related to each other than either of them are to someone from sub-Saharan Africa). So the frequencies of particular values for the markers will be different in different populations - perhaps 23 for the FGA marker is common in Egyptian people but not in Europeans. I think the markers used have been analysed across various modern populations of people, and the frequencies in each population are known.

So when you're doing DNA testing on a modern person you can make some judgement about what population group their ancestors came from based on the values you see at each marker. So this is what it means when it says that these results are 90% likely to belong to people of sub-Saharan African descent.

I'm not sure how relevant this is to the ancient Egyptian data - there isn't much data from the population or indeed from any people of that time period, so I don't know if the probabilities given will be valid or not.