A problem with the "Eddowes Shawl" DNA match

[Peter Griffith aka gryff]

I note that Mr Edwards' site [store and tour] has been very quiet.

Just looked at the news section nothing since Sept 8.

Funny that.

Well GUT as I posted earlier there was supposed to be a tour last night, and the "Halloween" tours start 25/10/2014

Halloween Tours

I'm not sure how many people can be handled on a "tour", but it does seem like "small potatoes" given all the interviews with the BBC and CNN (and others) that RE and Dr. JL were doing today.

cheers, gryff

[Panderoona]

I note that Mr Edwards' site [store and tour] has been very quiet.

Just looked at the news section nothing since Sept 8.

Funny that.

He has a new website thejacktheripperexperience and is offering halloween and bonfire night meal and tour at 25 quid.

[GUT]

I was looking at

http://www.jacktherippertoursandstore.co.uk/

Under "Ripper News"

[GUT]

Well GUT as I posted earlier there was supposed to be a tour last night, and the "Halloween" tours start 25/10/2014

Halloween Tours

I'm not sure how many people can be handled on a "tour", but it does seem like "small potatoes" given all the interviews with the BBC and CNN (and others) that RE and Dr. JL were doing today.

cheers, gryff

G'day gryff

Might be small potatoes, but you know "Look after the pennies and the pounds take care of themselves".

[Chris]

I believe this is the same entry percolated down to a more readable form.
It certainly contains the same pertinent (KJ748572(Russian)) information, halpogroup info and shows the same entry date.

KJ748572(Russian) FTDNA Haplogroup T1a1 07-MAY-2014
A73G T152C A263G 309.1C 315.1C 523.1C 523.2A G709A A750G A1438G
G1888A A2706G T4216C G4596A A4769G G4820A A4917G C7028T G8697A A8860G
T9899C T10463C A11251G G11719A C12633A G13368A C14766T G14905A A15326G
C15452A
A15607G G15928A T16126C A16163G C16186T C16188- 16193.1C C16294T T16519C

As described on " http://archiver.rootsweb.ancestry.co...-05/1400053213 "

That is very useful. I had started to pick out the sequence variations by hand last night (being too lazy to learn how to use a tool).

I don't think the presence of 315.1C there is significant in itself, as it's so common. However, there is one vital point that needs to be clarified. The book says that control samples from those who had been in contact with the shawl were found not to match the "Eddowes" sequence (for one particular short segment - other segments did match the controls, indicating contamination). But almost certainly most of the controls would have had 315.1C, and if so it must have been described as 315.1C (if he had also found 314.1C in the controls, surely Dr Louhelainen would have realised it wasn't so rare!). So obviously it needs to be checked that the controls still don't match when 314.1C is correctly described as 315.1C.

But what that information does allow us to do is look at how uncommon those sequence variations are, to try to get an indication of the significance of the "Kozminski" match.

It's not made explicit, but my reading of the book is that this match - like the "Eddowes" one - was based on a single short segment of DNA, presumably from one of the two hypervariable regions.

Looking at the sequence variations from the KJ748572 data you've posted, the second hypervariable region (1-574 in the reference sequence) contains only one that can be described as at all uncommon - the insertion of AC after 524 (here again there is an issue of nomenclature - the description on RootsWeb is different but equivalent), which is found in 3.4% of the sequences on EMPOP. Each of the others is found in 27% or more.

For the first hypervariable region (16001-16568), there are several variations with frequencies of 1.5% upwards in the range 16163-16193, but each of the others is found in at least 10%.

So - always assuming this is the right sequence in the GenBank database - I think the question would be which of the hypervariable regions the matching segment came from. If it came from the second it would probably not be a very significant match - particularly if it didn't include the AC insertion after 524.

[GUT]

That is very useful. I had started to pick out the sequence variations by hand last night (being too lazy to learn how to use a tool).

I don't think the presence of 315.1C there is significant in itself, as it's so common. However, there is one vital point that needs to be clarified. The book says that control samples from those who had been in contact with the shawl were found not to match the "Eddowes" sequence (for one particular short segment - other segments did match the controls, indicating contamination). But almost certainly most of the controls would have had 315.1C, and if so it must have been described as 315.1C (if he had also found 314.1C in the controls, surely Dr Louhelainen would have realised it wasn't so rare!). So obviously it needs to be checked that the controls still don't match when 314.1C is correctly described as 315.1C.

But what that information does allow us to do is look at how uncommon those sequence variations are, to try to get an indication of the significance of the "Kozminski" match.

It's not made explicit, but my reading of the book is that this match - like the "Eddowes" one - was based on a single short segment of DNA, presumably from one of the two hypervariable regions.

Looking at the sequence variations from the KJ748572 data you've posted, the second hypervariable region (1-574 in the reference sequence) contains only one that can be described as at all uncommon - the insertion of AC after 524 (here again there is an issue of nomenclature - the description on RootsWeb is different but equivalent), which is found in 3.4% of the sequences on EMPOP. Each of the others is found in 27% or more.

For the first hypervariable region (16001-16568), there are several variations with frequencies of 1.5% upwards in the range 16163-16193, but each of the others is found in at least 10%.

So - always assuming this is the right sequence in the GenBank database - I think the question would be which of the hypervariable regions the matching segment came from. If it came from the second it would probably not be a very significant match - particularly if it didn't include the AC insertion after 524.

G'day Chris

That post made my head hurt.

[Chris]

That post made my head hurt.

Sorry about that. If you don't want to see percentages, look away now.

These are what I got out of the EMPOP database for the frequencies of the variations in that Russian mitochondrial sequence:

Hypervariable region 2:
A73G - 76.0%
T152C - 27.0%
A263G - 98.8%
309.1C - 56.3%
315.1C - 99.2%
523.1C 523.2A - 3.41%

Hypervariable region 1:
T16126C - 12.6%
A16163G - 1.63%
C16186T - 1.51%
C16188- - 1.64%
16193.1C - 8.87%
C16294T - 10.4%
T16519C - 61.2%

If that is the right sequence, and if the match with the DNA from Aaron Kozminski's relation is based on just one segment from one of those regions, then it may be that there are no mutations involved that are at all rare. In which case it could be the kind of match that would be expected to happen anyway by chance.

For the "Eddowes" match, there were seven segments taken from those two regions, whose total length is 1142 positions. So the segments are probably quite short.

[GUT]

G'day Cris

But do I then combine those percentages.

So if we had

309.1C - 56.3%

and

C16294T - 10.4%

would it be about 5.5%

[Chris]

But do I then combine those percentages.

So if we had

309.1C - 56.3%

and

C16294T - 10.4%

would it be about 5.5%

It's a bit more complicated than that, because of the way the mutations are inherited. And we'd also need to take account of the fact that the figure would be less than 100% for other positions along the segment where mutations occur in the population. Fortunately the EMPOP search engine could calculate all that for us, if we knew where and how long the segment was.

The other point is that if the procedure was the same as for the "Eddowes" match, there would be up to seven segments, and therefore seven tries at a match. Even if the probability of success for each segment was only 5%, the probability of success in one of the seven might be more like 30%.

[mickreed]

G'day Chris

That post made my head hurt.

G'day GUT,

If you're going to face JL in the witness box, you must get a grip of this stuff. Otherwise I'm going to find another brief.

[Caligo Umbrator]

Hi Chris, Mick et al

I've set Aussie hours as a third clock on my PC so I can better track the different time zones on here.

I'm still trying to figure out how Hyper-variable regions work. The region 2 seems to have much lower percentages than region 1.

I've been checking all the listings and data sets they have at NCBI and, as yet, the one I linked to above is the only one that matches the criteria.
I can say now that I'm reasonably confident this is the same data set used by JL to match the 'AK' sample to T1a1.
It shows the correct haplotype, the correct stated country of origin and also fits into the time brackets as expressed in the book (I checked 4 months this side and over 18 months back on the other side).

And - APOLOGIES to Mick Reed for part of my earlier post - I may have seemed somewhat brusque or rude, especially regarding the link you posted.
I wasn't dismissing it.
I was juggling 3 things at home, suffering from sleep deprivation and having browser issues as well. Some of the frustration I was feeling may have come out in that post but I certainly didn't intend to direct any of it toward you or for it to read the way it does.
It actually is an interesting and useful page - fascinating to see all the variations and similarities that people have.

Yours, Caligo

[mickreed]

And - APOLOGIES to Mick Reed for part of my earlier post - I may have seemed somewhat brusque or rude, especially regarding the link you posted.
I wasn't dismissing it.
I was juggling 3 things at home, suffering from sleep deprivation and having browser issues as well. Some of the frustration I was feeling may have come out in that post but I certainly didn't intend to direct any of it toward you or for it to read the way it does.
It actually is an interesting and useful page - fascinating to see all the variations and similarities that people have.

Yours, Caligo

No worries, Caligo. I hadn't even noticed anything untoward. Maybe I'm half asleep as well. More likely I was worried that Gut might get me locked up given he's not really up to tackling JL in the witness box - he needs to get on top of those percentages in case JL brings a class action against us all.

Seriously, thanks mate. But I truly hadn't noticed anything. Don't worry.

[mickreed]

JL may be talking to the Finnish Press.

Part of a report on the Independent story, courtesy Google Translate:

Now Louhelainen is, however, already tired of the media blitz and the whole Jack the Ripper
….
Louhelainen said on Tuesday to Helsingin Sanomat.

"In fact, I'm trying to get back to a normal life, but it seems that 126 years ago the murders are more interested than ebola and submarines, which I really do not understand."

http://www.hs.fi/ulkomaat/a1413855388284

[mickreed]

More from Finland via Google Translate

- We stand by the conclusions presented in the book corner, Louhelainen says.

- We find, however, these allegations of possible errors. Nothing, however, change the fact that the DNA samples are compatible and their probability value is very high. The proof, for example, possession of the haplotyyppidata.

http://www.iltasanomat.fi/ulkomaat/art-1288753235986.html

The same article seems to blame the publishers for dumbing it down. The translation is so poor that it's not all clear, though.

[mickreed]

This one's quite funny

http://www.mhpbooks.com/spoilsport-scientists-ruin-recent-jack-the-ripper-book-revelation-with-their-stupid-bummer-science/