A problem with the "Eddowes Shawl" DNA match

[mickreed]

Well, I agree with that. Although, if they have made an error, then I think they will be in a tough position either way. It would be hard to say "Whoops, we made a mistake."

RH

Yes Rob, tough but vital. And I do agree any suggestion of fraud or deceit is absolutely uncalled for. However, if it can be proven that a mistake has been made, and I have absolutely zero doubt that it can be so proven, then not to face up to it would be very damaging.

The only possibility for redemption might be to release the actual data for examination. There may just be something in the data that is not in the book, but supports the claims made in the book.

[Monty]

The good Doc, looking kinda lost today, outside the White Hart.

He was having some photos taken, obviously for something promotional.

Monty

[mickreed]

The good Doc, looking kinda lost today, outside the White Hart.

He was having some photos taken, obviously for something promotional.

Monty

And this is the odd thing, don't you think?

He does seem to be playing a promotional role. That may be illusory but it could appear that way. If he is, then why is he?

[Monty]

There overheard conversation was telling Mick.

Apparently, he was filmed in Mitre Square yesterday.

Monty

[mickreed]

There overheard conversation was telling Mick.

Apparently, he was filmed in Mitre Square yesterday.

Monty

Hmmmm! Well, all I can say is that the higher you climb, the further you have to fall.

I am sure that it can be shown, beyond all doubt, that the material in the book regarding the rarity of the so-called 314.1C mutation is utterly mistaken. Unless there is something in the data that isn't in the book, then it's all built on sand.

Question is, how best to expose that mistake to the wider world?

[MrBarnett]

And this is the odd thing, don't you think?

He does seem to be playing a promotional role. That may be illusory but it could appear that way. If he is, then why is he?

It's just his good nature. He'll help anyone out, Scouse entrepreneurs, documentary producers... If you'd asked him for a quid to get a cup of tea, he would have given you his wallet.

Shame on anyone who thinks he might be in it for the dosh.

MrB

[mickreed]

Science backs Chris Phillips

What several of us initially thought was the only part of RE’s book that was possibly of interest - namely the suggested link of DNA on the shawl with the Eddowes family is, almost certainly, wrong. Just to recap, the book claims, seemingly in JL’s own words, that:

One of these amplified mtDNA segments had a sequence variation which gave a match between one of the shawl samples and Karen Miller’s DNA only; i.e. the DNA sequence retrieved from the shawl did not match with control reference sequences. This DNA alteration is known as global private mutation (314.1C) and it is not very common in worldwide population, as it has frequency estimate of 0.000003506, i.e. approximately 1/290,000. This figure has been calculated using the database at Institute of Legal Medicine, GMI, based on the latest available information. Thus, this result indicates that the shawl contains human DNA identical to Karen Miller’s for this mitochondrial DNA segment.

On the face of it, the claimed rarity of this ‘match’ does seem significant. But is it true? The claim began to unravel when a couple of people posted a reference to an article that suggested that the rare ‘global private mutation’ 314.1C was an ‘error of nomenclature’ and, as such, suspect.

Chris Phillips researched this, with some support from others, and he found that it was indeed an error. The rCRS has TCCCCCG for the positions 310 to 316. The meaning of 314.1C is that an extra cytosine (C) was inserted after the position 314 to give a string of six Cs rather than five, so that the zone shows TCCCCCCG. The same configuration, that is a string of six Cs, can be described as 310.1C, 311.1C, 312,1C, 313.1C, 314.1C, or 315.1C. Each of this will give a string TCCCCCCG, and it is impossible to tell the difference, because there is no difference. Because of this problem, the scientific community has agreed to standardise the nomenclature by using 315.1C to describe the string of six Cs rather than the five found in the rCRS. Indeed, the correct 315.1C applies to almost every European rather than the 1 in 290,000 claimed by JL and RE.

A lot of software, when a sequence containing an ‘error of nomenclature’ such as 314.1C is input, corrects this and replaces it with 315.1C. However the Haplogrep software chosen by JL and RE does not. What it does is to flag the problem as a ‘global private mutation’, which appears in bright red letters. The correct 315.1C is displayed as a ‘hot spot’ in green letters. The ‘global private mutation’ is Haplogrep terminology for an entry that does not appear in the authoritative Phylotree database. For it not to be found there, it must either be very rare, or a mistake. JL and RE seem to have preferred the very rare option. What it really means when that red flag labelled ‘global private mutation’ appears, is that the user should check their data. Had this been done, then the ‘mistake’ alternative would have been discovered.

A number of posters to this, and similar, threads took exception to Chris’s research findings, preferring to believe the DNA specialist rather than the non-specialist. Without the confirmation of qualified scientists, it was difficult to move beyond claim and counter-claim..

In order to move beyond this impasse, I tried to discuss the matter with JL but this was not possible. So I contacted the two bodies that are central to the discussion: Haplogrep at the University of Innsbruck, and Phylotree at the Department of Forensic Molecular Biology at Erasmus Medical Centre in Rotterdam. I included information, especially a book, that I'd previously posted here.

I asked them whether an entry such as 313.1C or 314.1C ought, as Chris had suggested, be corrected to read 315.1C, and if so, in the case of Haplogrep, whether it would help for an ‘error of nomenclature’ to be indicated in some other way. Here are their replies.

From Haplogrep:

Hi Mick!
You’re right – if it comes to these positions [313.1C and 314.1C], they should be put at the end of the C stretch [315.1C] – as denoted in the SWGDAM (under the FBI) Guidelines (please find it attached):
Rule 8 - Place indels on the 3’ end of the light strand. This will be available in the FASTA Sequence import that I’m currently working on. However if those are uploaded as variant files, no correction are currently made.
I can add an additional information in the new HaploGrep 2 version with very little effort, pointing to such common nomenclature errors. Positions around 310, 520 and 16189 to name the most prominent.
Thanks for this suggestion, really helpful – we often miss the forest for the trees.
Best,
Hansi

Note that Hansi says that Haplogrep does not correct where variant files are uploaded, and that something is going to be added to the programme in response to my suggestions.

And from Phylotree I got:

Dear Mick,

You are fully right and the book [Butler, Advance Topics in Forensic DNA Typing, p. 429] explains it well. 313.1C and 314.C are just the same as 315.1C as it is a stretch of Cs which becomes longer and you can position the extra C anywhere you want in the stretch. But the convention is to consider it as an insertion at the end (in DNA terms: at the most 3' position possible), hence 315.1C.
As the book says, the rCRS has a stretch of five Cs at 311-315. Almost everyone else has 6 Cs so 315.1C is an extremely common variant (in fact the rCRS is the odd one out here).

Best regards,
Mannis van Oven

So there we have it. Two of the chief scientists behind the software and databases upon which JL and RE seem to have relied are saying that 314.1C ought to have been entered as 315.1C and that the latter is extremely common rather than very rare.

I would suggest that this leaves the claims about the Eddowes DNA in tatters. There may be something else in the data that supports the claims made in the book, but the chosen data (314.1C) does not cut the mustard. Without seeing the entire DNA data files, it must be suggested that the claim of a very rare mutation in the DNA from the shawl and from Karen Miller does not stack up. I think we would all welcome full disclosure to see whether there is any real evidence that Eddowes family DNA is really on that shawl.

[Cogidubnus]

I think it begins to look pretty conclusive then Mick

Well done Chris, Debs et al

All the best

Dave

[Peter Griffith aka gryff]

In order to move beyond this impasse, I tried to discuss the matter with JL but this was not possible.

Mick, Would you care to elaborate on your efforts - what/how/why it was "not possible"?

And, by the way, a nice description of the problem and the responses you got from Haplogrep and Phylotree are indeed interesting.

Well done.

cheers, gryff

[Cogidubnus]

Hangs head in shame...

Well done Chris, Debs et al

Gosh how could I not have included Tracy who started it all off. Sorry!

All the best

Dave

[c.d.]

Thanks Mick for your post and all your work on this. If you could dumb it down for non-scientists that would be great. Does the shawl show a mutation that Karen Miller's DNA does not? Is that correct? Thanks.

c.d.

[John G]

Thanks Mick, this is excellent research. I think that Dr Jari and Russell Edwards' claims are now seriously undermined, particularly when you consider that Kosminski's haplogroup, T1A1, doesn't seem to be particularly rare either.

Cheers,

John

[tji]

Hi Dave

Gosh how could I not have included Tracy who started it all off. Sorry!

Thank you for the apology but it is really not needed. The acknowledgements really do deserve to go to Chris for his hard work and Deb's and Mick for their research.

Tracy

[lynn cates]

quickie

Hello CD. Permit me.

The upshot of Chris's findings is that about 99% of people have the "mutation." That likely includes you and me.

Cheers.
LC

[c.d.]

Ok. Thanks, Lynn. Got it.

c.d.