Jack the Ripper???

The DNA community had some exciting news this past week about the identity of Jack the Ripper, notorious serial killer of prostitutes in the Whitechapel area of London in 1888.  In total, there were 11 murders potentially linked to Jack the Ripper, with 5 being considered the most likely to be positively his victims.  He slit the throats of his victims, in some cases disemboweled them and mutilated their faces.


While there were many suspects and much speculation, the identity of the murderer was never established.  Among the suspects was one 23 year old Polish immigrant, Aaron Kosminski.  Aaron worked and lived in Whitechapel and was reportedly seen with one of the victims, but incriminating evidence was not given by the witness and he was released.  In 1891 he was committed to an insane asylum, probably a paranoid schizophrenic, where he eventually died.  He heard “solitary voices” and indulged in “unmentionable vices” which typically means activity of a sexual nature.

Last week, the British tabloid newspaper, the Daily Mail, ran a “world exclusive” article that Jack the Ripper has actually been identified as Aaron Kosminski utilizing DNA evidence found at the scene of one of the murders, that of Catherine Eddowes.

This was followed almost immediately by articles much more skeptical in nature, one in the Oregon Live and one by our own Judy Russell.

The reader’s digest version of the DNA part of the story is that a shawl was found with Catherine when she was murdered, although there is no evidence that the shawl was hers.  It’s believed that the killer left the shawl for some unknown reason.

The first problem with this story is that there is no proof that this shawl was indeed found with the body.  Catherine was so poor she reportedly hawked her shoes the night before, and the shawl in question was worth more than the shoes.  She has also just been released from jail for drunkenness before she was found murdered, and no shawl was mentioned by anyone.  Just the same, that doesn’t mean the shawl didn’t exist, and there is powerful DNA evidence, if it’s accurate, suggesting that this shawl was found exactly as stated, with Catherine’s body.

Russell Edwards purchased the blood-soaked shawl at auction, the shawl purportedly being found by a policeman the night of the murder and taken home to his wife, a dressmaker, who put it away unwashed.   Edwards hoped to somehow use the shawl to prove it was not only authentic, but to identify Jack the Ripper.

Edwards contacted Dr Jari Louhelainen, a leading expert in genetic evidence from historical crime scenes who combines his day job as senior lecturer in molecular biology at Liverpool John Moores University with working on cold cases for Interpol and other projects. He agreed to conduct tests on the shawl in his spare time.

Catherine’s DNA

He was able to extract DNA from some of the blood on the shawl and eventually managed to obtain mitochondrial DNA results.

Edwards managed to track down an individual, Karen Miller, who descends from the same matrilineal line as Catherine Eddowes, her three times great-granddaughter, and the mitochondrial DNA matched.  This is interpreted as confirming the identity of the blood on the shawl as that of Catherine.

Herein lies the second problem.

The article states that they “managed to get six complete DNA profiles from the  shawl” and that they were “a perfect match.”

I’m assuming, here, and I passionately hate to assume, because we all know what assume does…but I’m assuming that they are referring to mitochondrial full sequences here, all 16,569 locations on the mitochondria.  It would have been very helpful had they stated exactly what they tested.

They also don’t tell us what haplogroup they are working with.  If this is haplogroup H, it’s possible to have hundreds of “exact matches” because haplogroup H, itself, comprises almost 50% of Europeans today.  Of course, if they managed to sequence the entire mitochondria, the results would likely fall into a subclade, and some subclades are very rare, even within haplogroup H.

Because haplogroup H is so large, there is a great deal of diversity within H, and many of the subclades are small.  Furthermore, some people have no “unusual markers,” and those people tend to have many more matches than people who do have “unusual markers.”  Unusual markers are those mutations that have probably occurred in a family line and are not generally found in the majority of those of that particular subclade.

By way of example, here are the results from someone who is a member of haplogroup V, from eastern Europe.  They do not fall into a subclade of V and they have several extra mutations and one missing mutation compared to what is typically found in haplogroup V participants.


This individual has 3 full sequence matches, two of which are exact matches, but neither of those lead to the same ancestor.  This is a rather typical situation, not out of the ordinary.

The Ripper’s DNA

Another discovery on the shawl was that of semen, possible evidence of the Ripper himself.  They enlisted the help of Dr. David Miller who found surviving epithelium cells, a type of tissue that coats organs, in this case, thought to have come from the urethra during ejaculation.

Here a quote from Dr. Louhelainen about the DNA findings from these cells.

“Then I used a new process called whole genome amplification to copy the DNA 500 million-fold and allow it to be profiled.

Once I had the profile, I could compare it to that of the female descendant of Kosminski’s sister, who had given us a sample of  her DNA swabbed from inside  her mouth.

The first strand of DNA showed a 99.2 per cent match, as the analysis instrument could not determine the sequence of the missing 0.8 per cent fragment  of DNA. On testing the second strand, we achieved a perfect 100 per cent match.

Because of the genome amplification technique, I was also able to ascertain the ethnic and geographical background of the DNA I extracted. It was of a type known as the haplogroup T1a1, common in people of Russian Jewish ethnicity. I was even able to establish that he had dark hair.”

Here is the third problem.

This description seems to combine two types of sequencing.  Now, that’s not a bad thing, it’s simply confusing.  Based on the haplogroup of T1a1, we know that they sequenced mitochondrial DNA and that they did in fact manage to sequence it to the full sequence level.  How do we know this?  Because each mitochondrial haplogroup is designated by certain specific mutations.  In this case, the final 1 of T1a1 is indicated by location 9899 in the coding region of the mitochondria – so in order to designate this individual as a member of haplogroup T1a1, they had to sequence the coding region.  Again, we presume (the cousin of assume – with the same consequences) that they were able to successfully sequence the entire mitochondria.

Now for the fly in the ointment, I have not found this haplogroup in Russian Jewish people.  In fact, the clients who I have done DNA Reports for who fall into this haplogroup are not Jewish – none of them, nor do they have Jewish matches.  Neither does Dr. Behar identify this as a Jewish haplogroup in his founding mother’s paper.  Nor is this identified elsewhere as a Jewish haplogroup.  Of course, this Daily Mail article has no sources, so we can’t independently verify what was said, but it looks like this assertion of T1a1 typical of Jewish people may be in error.

However, from his discussion, we can also tell that additional sequencing has been done on the DNA retrieved, because you can’t determine traits like hair color without autosomal sequencing.  Therefore, if the descendant is truly related to Jack the Ripper, then at least part of their autosomal DNA should match as well, and that was not addressed.  If the autosomal DNA does not match, at least in part, then it calls into question the conclusions drawn by the mitochondrial DNA match.

We know that Kosminski was born about 1865 if he was 23 in 1888 when the crimes were committed.  The DNA matches a descendant of his sister.  Let’s assume, for purposes of argument that his sister was born about the same time.  And let’s use the standard genealogy generation of 30 years.  This means that the sister’s child was born in 1895, her child in 1925, her child in 1955 and maybe yet another child in 1985.  That’s a total of 6 DNA transmission events to a common ancestor, being the parents of the Kosminski siblings.  Therefore, Kosminski is the great-great-uncle to the child born in 1955.  Therefore, the individual born in 1955 should share about 6.25 of their autosomal DNA with Kosminski.  If they don’t, then there’s a problem.

If they do, then why didn’t the article tell us that.  This information would, in essence, seal the deal – well, assuming all of the other presuming is remedied.

Is It True???

First, let me state that in science, I’m always very, very skeptical of publication via newspaper or internet, especially publication via tabloid.  This has been fraught with problems.  Debbie Kennett has covered this repeatedly on her blog.  Another example is the announcement of  Pict DNA being identified – published and never proven.  I know of other cases in which DNA evidence is intentionally twisted, inaccurately, to fit the intentions of the publishing entity.  So, yes, I’m a rabid skeptic without provable evidence.

I want to see this assertion go through the verification process with a second, reputable, lab.  By reputable, I mean one not associated with any of these other questionable assertions.  Then, I’d like to see the results published in an industry accepted journal.  Yes, that takes time, and yes, there are questions to answer, but the resulting paper carries with it credibility that is impossible to obtain otherwise.  Unfortunately, publishing results in a tabloid paper immediately causes me to question why they would have made that choice if they had solid proof.

Ok, now that I’ve said that, I want to address the question at hand.  Is it true?  Might it be true?

I’d like to make two points.  First, while I used random examples of mitochondrial matching, this isn’t a random situation.  This is a known individual in both cases, with known and I’m assuming, provable, genealogy to both Catherine Eddowes and to Aaron Kosminski.  We’re not looking at random matches here and we’re not looking for a common ancestor.  We know who the common ancestor is in both of these situations and we’re looking for matches to confirm that identity.  This, by the way, is exactly how our armed forces identify remains of soldiers and repatriate them to the family.  This uses the exact same premise – that we’re not looking for random matches, but for a match with a known family member of known provenance – with possibly, hopefully, family line mutations.

Now, let’s use a bit of math, which is sometimes, but not always, my friend.

I’m going to use two examples, haplogroup T1a1 and haplogroup J1c2f because it’s mine and I have easy access to those results.  We know that the mitochondrial DNA attributed to Kosminski is T1a1 and we’ll just let mine stand in for Catherine Eddowes.

In the Family Tree DNA data base, haplogroup T represents 8.06% of the participants and haplogroup J, 7.77%.  As of September 8, they have a total of 43,329 full sequence mitochondrial DNA results in their data base.  I calculated the number of members of each haplogroup based on that percentage and then I checked the corresponding DNA project at Family Tree DNA.  Then I checked to see how many occurrences of the subgroups of J1c2f and T1a1 were found and calculated the percentage of haplogroup J and T they represent.  The total subgroup percentage is the percentage of J1c2f and T1a1 of the entire FTDNA full sequence population.

  % FTDNA # Members Hap Project # #J1c2f/ T1a1 % of  Hap Proj Total % subgroup
J1c2f 7.77 3336 2165 6 .2 .01
T1a1 8.06 3492 673 52 .73 .12

London’s population was estimated to be 1 million in 1800 and 6.7 million in 1900, so let’s use the figure of 6 million for 1888 as an estimate.

Of 6 million people, you would expect to find 600 people carrying haplogroup J1c2f and 7200 people carrying T1a1.  Therefore to find two of those individuals whose DNA is found on the same scarf, who have a forensic tie, or a suspicion of a tie, is astronomically small.

If math is my friend today, we would multiple values of each haplogroup in the population together to find the odds of finding both in one place.

That would be .0001 times .0012, which equals 1.2e-7 which means, 0.00000012, in other words, about one in 1.2 billion.  The population of the world in 1875 was calculated to be about 1.3 billion

So, assuming their work is accurate, and assuming that this isn’t a huge elaborate hoax, it’s very likely accurate, and Jack the Ripper is very probably Aaron Kosminski.

Where’s the Beef???


Remember the old Wendy’s refrain, “Where’s the Beef?’’

Well, I want to believe this story, especially since it’s such a feel good fairy tale story involving a Jack the Ripper hobbyist and DNA, of course.  But I’m really left waiting for some kind of corroboration.  Was it Carl Sagan that said “extraordinary claims require extraordinary evidence?”  Well, they do and I really hope the authors will subject their findings to peer review and authenticate their claims.  If this isn’t true, it’s a hugely elaborate and well-planned hoax perpetrated probably to sell a resulting book or movie which should, if that is true, be named “Jack the Ripoff.”

I want this to be true, and I want the authors to make a believer out of me.  I want no presumes or assumes left standing.  So….where’s the beef???

29 thoughts on “Jack the Ripper???

  1. A very interesting and detailed response, Thanks. Learned much about using DNA. I have an ancestor who has defied 30 years of research so had a male cousin do a DNA test and I have found nothing that ties in with the ancestor. Do I assume the ancestor was using a made up name? The family story of his relationship to a known person (General William Tecumseh Sherman) can not be proven. Anyway, I enjoyed your article and look forward to more. bfhodgkin@gmail.com

  2. Thanks for the really interesting analysis!

    There’s some other issues, assuming that the DNA work is correct. Kosminski was known to frequent local prostitutes, and may have been a client of Eddowes. Just because his DNA was found on a shawl she had been in contact with, only meant that he too had been in contact with the shawl, not that he was the killer. If his blood had been found on the scarf (as many knife attackers also cut themselves in the process) that would have been more significant than just semen. Eddowes could have been killed by the Ripper after Kosminski left his DNA on the shawl. I would be interested in a history of the type of shawl that was found, was it a common or unusual design, was it the sort a Jewish immigrant might be in possession of, was its cost within Kosminski’s means? If it was an expensive shawl, would it make sense for Kosminski who wasn’t wealthy, to have left it? Could Kosminski or another client have paid Eddowes with the shawl, etc?

    Additionally, “two of Eddowes’s descendants are known to have been in the same room as the shawl for 3 days in 2007,” [Burgess, Kaya (8 September 2014). “DNA row over ‘proof’ Aaron Kosminski was Jack the Ripper”. The Australian.], so contamination is certainly a possibility.

    • So true and agree with you! More research and DNA need to be done with
      garments of other victims. One “possible” result – does not answer who was Jack the Ripper. I think that night that this victim died, where is also
      another earlier possible victim. ?2 murders in one night by the same person. Are there more garments and/or sexually assults to review. Please keep up posted. I do understand- Newpapers, want to sell, make money with their postings; giving readers little in actual facts in the science of the findings.

    • Another article said the shawl had a MichaelMas design, a Christian holiday. Apparently, it matched the date of her murder. No Jew would have something like that, though it’s possible a murderer might either take what was on hand or might try to leaving confusing clues.

  3. Great post Roberta!! I find it very interesting. Then I thought “oh no, could I be related to Jack way back in the primordial soup of the T2 haplogroup? I am T2b3 European lines.


  4. It kills me that they’re balancing this whole thing on a DNA sample of semen. I don’t mean this in a judgmental way (truly), but let’s be real: Jack the Ripper’s victims were prostitutes. Prostitutes come in contact with multiple semen donors. It’s ridiculous to assume that the only semen that got near a victim of Jack the Ripper was that of Jack himself. It’s not like she would have been taking that shawl to the laundromat every few hours.

    Also, I’m a T1a1, and the whole Russian Jew thing threw me for a loop. My understanding was that T1a1 led mostly to people with British Isles or Scandinavian ancestry (which is consistent with my maternal line’s known origins). I think that’s a particularly weak part of the argument.

    All that said, I’m sure the guy will sell a whole bunch of books, which is usually the whole point. I hope Kosminski IS the killer, because if not, a grave injustice is being done to his memory.

    • I am T1A1j, and I’m Jewish for more than 7 generations of my mother’s families. They were mostly Russian Jews. Some Eastern Europe as well.

  5. So where has that shawl been all these years, floating about in someone’s cupboard..! How come it wasn’t in the Police Museum at Scotland Yard. It has to be highly suspect due to the contamination of evidence over the intervening years. I agree with Kathryn, as I also thought that Kosminski could have visited with Eddowes, and she could have met with the Ripper later. Some years ago a documentary claimed that it was almost certain that the Ripper was an American Doctor, who had returned home. Also that the way the women’s organs had been removed pointed to someone with medical knowledge.
    I have no confidence in the Daily Mail either, a gutter rag as far as I am concerned. Your points about the DNA not being of Jewish origin, and their claim that Kosminski, a Jew, is guilty, needs to be confirmed or rejected by the highest authority possible, especially with the present rise in Anti-Semitism. I am suspicious of their (Daily Mail) motives.

  6. Thanks for writing this – I too was skeptical when I saw the claim and where the claim was posted – in a rag! I am hoping Debbie Kennett will also write something. I was disappointed with (thus far) the brief comment on Dieneke’s Anthropology Blog as he usually is far more critical – maybe waiting for the details?

  7. Many thanks for a truly fascinating read!

    My maths probability is a little rusty, but unfortunately I think you’re falling foul of Prosecutor’s Fallacy when you multiply the two probabilities together. Two probabilities can only be multiplied if they are truly independent and uncorrelated of each other (e.g. tossing a fair coin). For example, see the Sally Clark case for an example of an unsafe conviction based on an expert witness multiplying probabilities.

    For example, if 7,200 men (I’m assuming you meant men, rather than both sexes?) in London were carrying T1a1 and Kosminski was one of them, that means the chance it being Kosminski’s semen is 1 in 7,200 — I’m assuming here this group is just men and that everyone in that group had an equal opportunity to leave semen on the garment (ewww!). You would then need to prove beyond reasonable doubt that it couldn’t belong to any of the other 7,199 men.

    Similarly, with the provenance of the garment (apparently a table runner rather than a shawl) in such doubt, unless you could prove the garment was taken from the scene and was soaked with Eddowes’ blood, you would need to show that the blood couldn’t have come from any other person in that haplogroup.

    I’m probably committing Defense Attorney Fallacy here, but I think the 0.00000012 figure is more representative of the chances of the blood and semen actually belonging to Eddowes and Kosminski respectively. In other words, it’s more much likely that they belong to other unknown individuals.

    • Actually, the 7200 is total males and females, so if you would presume half of those are males, it drops to 3600. The 1 in 1.2 million figure is the probability that both of those DNA haplogroups would arrive on that shawl without any relation to each other – by happenstance, independently – as some fine defense attorney would of course claim:)

      • Thanks, Roberta!

        It’s definitely a fascinating story 🙂 The garment has been examined and handled by many people over the years and, according the an expert at the Victoria and Albert Museum, is apparently a design that wasn’t even produced until the early 1900s(!)

  8. Thx for a good overview. Now, although the researcher id a Finn (eorking kn the UK) I fon’t know him and sll I knoe is based on newspapers and Internet. In Finnish newspapers Dr Louhelainen has been interviewed perhaps in more detail. Based on those articles in Finnish one special issue in methodology is the use of laser microscopy which has been used nearly just for cancer tests. This means that just one cell can be taken for DNA analysis, and no liquid is needed for that. The normal methods would demand liquid to get DNA out of a cloth etc and all kind of DNA (contamination) can come along.

    Besides semen, there was obviously e.g. some kidney cells from the victim. Firs with Jack the Ripper’s habits.

    A couple of top-class Finnish researchers have commented they are cautiously optimistic.

    Hopefully we’ll soon see a scientific publication explaining the method.

  9. A great piece of clinical scientific analysis Bobbi and one everyone has to respect! As an aside, when I did my piece on Jack the Ripper as a teenage London schoolboy (in the days before DNA), the injuries to the bodies were adjudged to have been conducted by someone of immense physical strength, but oddly, with some knowledge of anatomy. This latter facet was the only thing I could find to argue against Kosminski (although he certainly had the physical strength). My only other shortlisted candidates were: Sir William Gull (Queen Victoria’s physician, for which there is compelling evidence to support his presence in Whitechapel at the time of at least two of the murders, and who was witnessed to having “insane rages” when dealing with some patients); and George Chapman, who ticks all the boxes excepting that he poisoned his three wives. Mutilation was not therefore his modus operandi!

  10. I’ll be curious to see how this turns out, but as a T1a1 I’ve gotta say that my understanding is this (pretty common) group goes all over northern Europe and northwestern Asia, simply being T1a1 doesn’t reveal much. My mother has 70 exact matches and the ancestry listed for them includes Ukraine, Norway, Finland, Sweden, Germany, Belarus, England, and the US (most, alas, don’t list ancestry).

  11. If it is a hoax to sell books, then why did they finger Kosminski as the perp? Was there no suspect more intriguing? If I am hoaxing this, I make the killer either bigfoot or a time-traveling Lee Harvey Oswald.

  12. The haplogroup mistake is easy to make as there is a Y-haplogroup of the same name – which has those characteristics which Dr. Louhelainen says. It looks like the books has been rushed to print and anyone could make that kind of mistake. Unless he means Y-haplogroup in which case he statement is correct.

  13. I would like to refine the T1a1 estimate.

    You have listed 52 T1a1 testers out of 43329 total FMS. If I understand that correctly you got those numbers from those who signed up for the T1 project. 0.12%

    I am T1a1 and I actually have 213 T1a1 matches at FTDNA. So thats about quadruple what you estimated or 0.49%. Altho I am not sure if this represents the total T1a1 population at FTDNA or just those who matched me (usually at 2-3 steps)

    This is roughly consistent with my results at 23andMe where I have 12 T1a1 matches out of 993. So that’s over 1%. Altho since these are my (remote) relatives, perhaps its skewed slightly in favor of my mtDNA haplotype.

    This article suggests that T1a1 is about 1.6% of the Western European population There are two or three relevant paragraphs in the middle of the paper.
    This article also implies that T1a1 can be found in other regions too, sometimes at slightly higher percentages.

    Overall I agree with your premise that the chances of finding two scarce haplotypes on the same scarf are astronomically small.

    btw, I am a true amateur on this subject, so I welcome any “refinements” of my comment. And its not that I really seek a distant connection to Jack the Ripper.

  14. Slight edit to the above.

    I have 213 FTDNA T1a1 matches at FMS. But I actually have 277 T1a1 matches if I include lower levels. And my guess is that there are others that I dont see because I do see a T1a1 at Gedmatch originally from FTDNA and she personally is not on my official FTDNA match list.

    • I can see from the FTDNA match list that I am 2 steps from Mr Abrams. I have the root form of T1a1, that has not mutated for 6997 years according to Behar,and I am Jewish. There are four persons, including me, of Jewish ancestry as far as I know out of 70 exact matches on FTDNA. They are from western Ukraine (2), south Belarus and Moldova. There is another Jewish person one step away from me, with a comparatively recent mutation (only 1,000 years ago?) from southern Ukraine (Volynia).
      There is however another group of Jewish descent described by a different haplogroup T1a1j with the additional mutation A8530G. This is an ancient mutation, dated by Behar as occuring 5222 years ago. I was in correspondence with a member of this group before we tested on the coding region after which he was removed as a match, since he was placed in a different haplogroup. There are at least five Jewish members of this haplogroup tested by FTDNA, at least equalling the non-Jewish members (Of course I have no up to date information). Two are from Belarus, one from Hungary and one from Turkey, of Sephardi origin in the male line, but possibly Mizrachi (Romaniot) in the female line. This member differs on HVR1 with my correspondent of Belarus origin, suggesting an ancient split (2,000 years). This further suggests an origin in the Middle East and immigration to eastern Europe only in historical times.
      This is against the understanding of all researchers except Dr. Pike. They say T1a1 originated in Europe in the Neolithic period or before. Perhaps a T1a1j woman returned to the Middle East 3,000 years ago, just as Pala suggests that the Morrocan and Turkish exact match with myself (the root form) are the result of a spread from Europe.
      Anyway to sum up from my microcosmic viewpoint, T1a1 is rare but present among Jews, but not known previously to me in central Poland, but much farther south and west. The flight west of Jews after the Chmielnitzki massacres of 1648/9 certainly reached the area of modern Poland, so it may not be be so surprising. But was Kozminski T1a1 or T1a1j? Presumably he did not have A8530G if he described only as T1a1. If he was T1a1j it one could stretch the truth and say it is a common haplogroup among Jews, at least as common as among non-Jewish members of the haplogroup, but that would not apply to T1a1.

  15. Pingback: DNAeXplain Archives – General Information Articles | DNAeXplained – Genetic Genealogy

  16. If Aaron Kosminski’s modern female relative carries T1a1 and she is Jewish, descended from the same female as Kosminski, then he had T1a1 without a doubt. This mitochondrial haplogroup is not very common among Eastern European or any Jews. Since the semen belonged to a male, of course, the odds that it was deposited by Aaron Kosminski are very good. He was the one who frequented the area in which Eddowes could be found, as well, was there consistently, which gives him an edge over British males of Viking ancestry with T1a1 who might have dropped by. It doesn’t prove Kosminski the murderer, but certainly keeps him a suspect.

  17. I don’t really know where to turn but I stumbled across this blog and this was the most scientific I have read about the subject so figured it was worth a shot: I think there is a chance my grandfather was the ripper.

    • Did your grandfather have any siblings who are living who could DNA test? If your grandfather had sisters who had children, those children could test too. You’d want a mitochondrial DNA test from Family Tree DNA to see if they carry the same haplogroup. Why do you think he may have been the ripper?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s