DNA and the law

[AGRBear]

I think there is a misunderstanding occurring.

The thread is about DNA/mtDNA in the courts.

It is my understanding from all the lawyers with whom I've had conversations that in the case of proving AA was FS that they would, along with the DNA/mtDNA need to provide the jurists with additional evidence. Why? Because the lawyers, who's job it is to prove AA was not FS, would present a very strong case and would probably accomplished the task of placing doubts about the DNA/mtDNA.

I merely took out of the speculations that anyone at the hosptial, Dr. Gill, Ginther and other scientists were involved in any kind of wrong doing. Instead, I was dealing with the time period between 1976 up to the time the samples of the intestines identified as being AA's became part of a court case when people, who were outside of the hosptial and labs had motives, opportuntities and money to support unlawful acts.

This is all that I'm trying to convey.

I am not here to try and convince you that AA was FS. I am merely trying to explain that emotions, logic and your own beliefs doesn't fly very far in a court of law when you dealing with 11 people and the 12th being Bear on the jury.

There really isn't much more I can post, unless, you really do want to go back through all the evidence for all those concern. But not here. This would need it's own thread.

AGRBear

[Elspeth]

Originally Posted by Michael HR

I thought the mtDNA from Carl Maucher was a match to that from Anna Anderson and would suggest that Anna Anderson was a member of his family. If she was there is I assume only one member she could have been.

The way you wrote your answer suggests that even if there is a mtDNA between AA and Maucher match this would not necessarily mean that the two test subjects are related to each other and if that were true then the match between Empress Alexandra and the children to Prince Philip might mean that they were not related?

Why would a macth between AA and Carl Maucher not mean they are related as this relationship with mtDNA seems to be the whole basis of the ID of Anderson and of course the Imperial Family with thier mtDNA?

Michael

mtDNA isn't the same as a genetic fingerprint; it isn't unique to a particular individual. There are different mtDNA patterns, some more common than others, so it somewhat depends on how common the sequence is as to how certain you can be that you're dealing with actual relations. This is what the Gill paper said about the mtDNA pattern from the Anna Anderson samples and from the Carl Maucher sample:

"To assess the strength of the evidence we compared the DNA profile from the Maucher and putative Anna Anderson sampels with over 300 Caucasian sequences in published and unpublished databases. We did not find a match, which suggests that the DNA sequence is rare. Assuming that the databases we have used are representative of European Causasians, the chance of finding matching DNA profiles if Carl Maucher and Anna Anderson are unrelated through the maternal line is less than one in 300."

That doesn't rule out the possibility of a coincidental match, but it makes it unlikely. However, the fact that the Anderson samples didn't match the Prince Philip sample means that she isn't related to Prince Philip through the maternal line, which she would have been had she really been Anastasia.

[BorisRom]

Originally Posted by Elspeth

Well, it doesn't matter how many or few people have congenital hallux valgus, does it, if the two ladies concerned had it to different degrees. All the statistics in the world aren't going to help you if the two individual cases aren't the same. Severe hallux valgus isn't necessarily the same for two people. <...>

Elspeth,
Reliability of DNA-test=1:N is meaning, that DNA-tests of N persons will show, that one of them is a relative of test-person. It is properties of work with the limited databank of DNA (the databank does not include DNA of all people on the Earth). How much I know, in 1990 reliability of DNA-tests was no more 1:6000. It means, for example, that if in 1990th years the genetics would compare my DNA to DNA of 6 000 000 Portugueses (for example) - in this case 1000 Portugueses would appear as my relatives. I think, that any Portuguese is not my relative actually (unfortunately ).
On the other hand, if I had 6000 relatives (for example) - in this case DNA-tests (in 1990th years) would show, that one of them is not my relative (in spite of the fact that he is my relative).

Now we shall consider medical statistics of C-HV (we shall assume for simplicity of comparison it=1: 18 million). AA said, that she is ANR. Both had C-HV. Only one of 18 million women could say it. DNA-tests (with reliability 1:6000) would apply it (AA=ANR) to 3 thousand women (from these 18 million)!!!
The medical statistics of a heavy case of HV (1:13000) would show, that only one of 13 thousand women could apply to be ANR, while DNA-tests (with reliability 1:6000) could apply it to two women (from these 13 000).

Thus, DNA-tests with reliability less than medical statistics of congenital HV (or at least of heavy HV) cannot be considered more preferable -
at least in such complex case as history of AA-ANR.
...At least in such complex case as history of AA-ANR...
.

[Elspeth]

Reliability of DNA-test=1:N is meaning, that DNA-tests of N persons will show, that one of them is a relative of test-person. It is properties of work with the limited databank of DNA (the databank does not include DNA of all people on the Earth). How much I know, in 1990 reliability of DNA-tests was no more 1:6000. It means, for example, that if in 1990th years the genetics would compare my DNA to DNA of 6 000 000 Portugueses (for example) - in this case 1000 Portugueses would appear as my relatives. I think, that any Portuguese is not my relative actually (unfortunately ).

Where are you getting these numbers?

And why are you even worrying about them? If mtDNA samples from two people don't match, they don't match. Statistics are irrelevant.

[BorisRom]

Elspeth.
Numbers of reliability of DNA-tests = 1:6000 (in 1990th years) is the most high that I saw - “1:6000” is in “FOTR” (by Greg King and Penny Wilson, - page 757 in Russian edition).

[AGRBear]

Originally Posted by Elspeth

mtDNA isn't the same as a genetic fingerprint; it isn't unique to a particular individual. There are different mtDNA patterns, some more common than others, so it somewhat depends on how common the sequence is as to how certain you can be that you're dealing with actual relations.

...[in part]....

I thought the combination of the DNA and the mtDNA did give us the genetic fingerprint and that is why they believe the male and female remains in the two pits found a year ago July are siblings and children of Nich. II and Alexandra. And that the markers would be nearly the same but not exactly so they will [so I'm told but have not yet seen the results of this] be able to prove there are four grand duchesses and not just three.

These same markers were used to compare Prince Philip's, the remains believed to be three grand duchesses, Alexandra and Nich. II's, plus the blood tests of Carl Maucher's. Marg.'s and the intestine samples which many believe were AA's. Other tests included hair, blood samples found on slides, syringes, etc..

Just using the sample of the intestines, Dr. Gill and Dr. Ginther agree that the intestines believed to be AA's show a close match with Carl Maucher's. It is likely that they are related when using the intestine and hair samples. This relationship could mean Carl can be the nephew or a relative. Meaning AA and Gertrude, Carl's grandmother, could be sisters or first cousins or second cousins up to 25th cousin. OR, the match could be just a coincidence.

In Dr. Gill's data base at that time had only 300 samples of Caucasian Europeans. Does not mention Russian samples. Since then, I'm sure these numbers have changed.

I always get itchy around numbers [stats] because the conclusions by individuals/groups usually depend upon who they are and why the numbers are gathered. (Living in Cailfornia and seeing all kinds of number floating around by the Dem. and the Rep. Parties during political campaigns, I could provide you with great examples of what oppositions like our Dem. and Rep. can do with their numbers.)

Added to this is the problem that no one has found the birth certificate and or baptismal record of Gertrude, nee S., Ellrich's with the other records of FS, Felix and the other siblings. So, we're not even sure that FS and Gertrude had the same mother and/or father.

According to Richard Schweitzer, a copy of a birth certificate was sent to him but it is without any kind of official seal. Nor do we know who found this information and where it was registered. Therefore, doubt circles this copy and cannot be presented, at this time, with any kind of official backing. I assume he's looking into this farther for verification of some kind. When I was looking, I found about 15 females around Posen with the same name [spelled a variety of ways] but none had the names of the parents of FS.

AGRBear

[Elspeth]

Originally Posted by BorisRom

Elspeth.
Numbers of reliability of DNA-tests = 1:6000 (in 1990th years) is the most high that I saw - “1:6000” is in “FOTR” (by Greg King and Penny Wilson, - page 757 in Russian edition).

Do you mean the quote (from page 445 of the English edition that I'm looking at):

"The sequences derived through this method [DNA splicing] matched that provided by the duke of Edinburgh, itself considered rare in the early years of genetic research when the testing was conducted: statistical estimation placed the likelihood of finding the same genetic profile at 1 in every 6,000 European Caucasians."

Because that isn't talking about the reliability of the test in the sense of whether the test is giving a meaningful match, it's talking about the likelihood that an mtDNA match indicates actual relatedness.

Which brings me back to the point I was making before. Since these tests aren't genetic fingerprints, a match doesn't imply a unique identity. That's well known, and the authors of the Gill paper have said as much. However, a mismatch rules out relatedness through the female line. That's why the authors of that paper were so much more certain that Anna Anderson wasn't Anastasia than they were prepared to claim that she was Franziska Schankowska.

The 1 in 6000 stuff refers to matches, not mismatches.

[Elspeth]

Originally Posted by AGRBear

I thought the combination of the DNA and the mtDNA did give us the genetic fingerprint and that is why they believe the male and female remains in the two pits found a year ago July are siblings and children of Nich. II and Alexandra. And that the markers would be nearly the same but not exactly so they will [so I'm told but have not yet seen the results of this] be able to prove there are four grand duchesses and not just three.

No, they don't give a genetic fingerprint. A genetic fingerprint would use so many markers that the likelihood of a perfect match being coincidental would be essentially zero. As far as I know, and I'm sure you know I'm not an expert, you wouldn't use mtDNA in a genetic fingerprint, just nuclear DNA. The point of a genetic fingerprint is to take DNA samples from two sources (such as a murder suspect and the crime scene) and compare them to see if they're from the same person. It can also be used to establish paternity if the DNA pattern of the alleged father matches enough of the DNA sequences of the child, making allowances for the fact that the child will have inherited half his DNA from his other parent.

In this case you don't have DNA from the two people concerned, or DNA from a parent or a sibling, you're just dealing with great nephews and second cousins or whatever. The DNA itself was pretty badly degraded and I don't even know if there would have been enough of it for a full fingerprint analysis.

These same markers were used to compare Prince Philip's, the remains believed to be three grand duchesses, Alexandra and Nich. II's, plus the blood tests of Carl Maucher's. Marg.'s and the intestine samples which many believe were AA's. Other tests included hair, blood samples found on slides, syringes, etc..

If I remember right, they were only using five or six markers, which is fairly standard in this sort of test. I think a full nuclear DNA fingerprint analysis uses something like 20. It's been a while since I read about this in any great detail so I could be misremembering.

Just using the sample of the intestines, Dr. Gill and Dr. Ginther agree that the intestines believed to be AA's show a close match with Carl Maucher's. It is likely that they are related when using the intestine and hair samples. This relationship could mean Carl can be the nephew or a relative. Meaning AA and Gertrude, Carl's grandmother, could be sisters or first cousins or second cousins up to 25th cousin. OR, the match could be just a coincidence.

Yep.

In Dr. Gill's data base at that time had only 300 samples of Caucasian Europeans. Does not mention Russian samples. Since then, I'm sure these numbers have changed.

They may well have. At this point my guide would be the fact that Dr Ginther said three years ago that the Gill results were generally believable; he didn't say that they were probably invalidated by larger databases of reference samples.

I always get itchy around numbers [stats] because the conclusions by individuals/groups usually depend upon who they are and why the numbers are gathered. (Living in Cailfornia and seeing all kinds of number floating around by the Dem. and the Rep. Parties during political campaigns, I could provide you with great examples of what oppositions like our Dem. and Rep. can do with their numbers.)

Indeed, and this is why you'll never see me claiming that the match with the Maucher sample indicates anything more than a high likelihood of relatedness, which is a lot different from a certainty.

However, the fact remains of the mismatch between the Anderson sample and Prince Philip's sample. In the face of a mismatch, statistics are irrelevant. If they don't match, they don't match. The rate of mutation of mtDNA through the generations is too low for people of such similar generations as Anna Anderson and Prince Philip to be related through the maternal line in the face of that sort of mismatch.

[BorisRom]

Originally Posted by Elspeth

Do you mean the quote (from page 445 of the English edition that I'm looking at): "The sequences derived through this method [DNA splicing] matched that provided by the duke of Edinburgh, itself considered rare in the early years of genetic research when the testing was conducted: statistical estimation placed the likelihood of finding the same genetic profile at 1 in every 6,000 European Caucasians."

Elspeth,
yes, you are right:
the likelihood that an mtDNA match indicates actual (valid) relatedness is 6,000:1 (in 1990-th years). Thus: one mistake on 6,000 tests.
Hence, 3,000 women from every 18 millions could say they was ANR (actual relatedness) through DNA-tests in 1990-th years – if all 18 millions would test on DNA (mtDNA mutch).
Now let me consider medical statistics of C-HV (we shall assume for simplicity of comparison it=1: 18 million). AA said, that she is ANR. Both had C-HV. Only one of 18 million women could say it. DNA-tests (with reliability 1:6,000) would apply it (AA=ANR) to 3 thousand women (from these 18 million)!!!
The medical statistics of a heavy case of HV (1:13,000) would show, that only one of 13 thousand women could apply to be ANR, while DNA-tests (with reliability 1:6,000) could apply it to two women (from these 13,000).

[Elspeth]

What makes you think the 1 in 6000 is referring to mistakes? It isn't the reliability of the test they're talking about, it's the probability, since an mtDNA profile isn't unique, that a match would be coincidental rather than indicating relatedness. The probability would alter depending on the rarity of the sequence in question.

[BorisRom]

Elspeth,
I mean a mistake that a match would be coincidental rather than indicating relatedness - namely it.
My example illustrates it (3000 coincidental on 18 million of DNA tests)

[Anna was Franziska]

Just for the record here, the DNA numbers presented in "FOTR" have been questioned by many DNA knowledgeable people on another board and have not been proven to be accurate. One of the authors has questioned the authenticity of the bones and supported the Knight paper.

http://peterkurth.com/RUSSIAN%20FORENSICS%20TEAM.htm

and here is the rebuttal explaination

http://www.facesofrussia.org/index.p...&id=3&Itemid=1

(for background informational purposes only)

[Elspeth]

Originally Posted by BorisRom

Elspeth,
I mean a mistake that a match would be coincidental rather than indicating relatedness - namely it.
My example illustrates it (3000 coincidental on 18 million of DNA tests)

The chance that a match would not indicate relatedness varies depending on how common the particular mtDNA sequence is.

[BorisRom]

Elspeth,
Well. In what borders this chance varies? (depending on sort of mtDNA sequence):
1:100? 1:1,000? 1:6,000?
If I understand correctly, 1:6,000 was the best (highest) chance (in 1990-th years).
Do you know?