The origin and migration of domestic cats: a genetic study

I think about fifty people sent me articles about a new genetic study of domestic cats and their ancestor, Felis silvestris—an analysis published in a paper in Nature Ecology & Evolution by Claudio Ottoni et al. Thanks to all for calling this to my attention, as it combines two of my favorite subjects, cats and genetics; but excuse me if I can’t thank you all by name.

The reference and free link to the paper (if you have “Unpaywall”) is at the bottom, as well as a link to the study’s supplementary material. The paper was also summarized in articles in The Guardian and in a Nature News and Views piece, and got tons of attention in the press because, well, cats.

In truth, the results can be summarized briefly; they’re a bit surprising but not earthshaking. First, if you want a video presentation and don’t want to read this whole post, just watch the 3.5-minute Nature synopsis below:

The authors looked at the mitochondrial DNA of 352 ancient cats from 30 archaeological sites, with samples taken from teeth, skin, and hair. They also looked at recent museum specimens of the five known subspecies of F. silvestris: the subspecies names are silvestris, lybica, ornata, cafra, and bieti. Here are their distributions with the numbers corresponding to the 30 archaeological sites sampled:

It’s been known from previous genetic studies that domesticated cats came from just one of these subspecies F. silvestris lybica (FSL), although modern housecats in Europe hybridize, and thus get genes from, the European subspecies F. silvestris silvestris (FSS). We also know, from remains of a cat associated with an ancient burial in Cyprus, that cats were at least semi-domesticated by 10,000 years ago, though they were probably not pets but living in association with humans and used for controlling rodents (there was agriculture by then, and stored grain needed protection). Here’s a rather skinny FSL, showing that the wild species is a striped (“mackerel”) tabby cat, but what’s shown below is a real wild species:

What the new study found, though, was that, using easily extracted mitochondrial DNA (mtDNA), there are five genetically distinct groups, or “clades”, within FSL. Using the distribution of these clades from dated cat remains from known locations showed how where cats were domesticated and how they moved about with the help of humans (probably largely on ships, which also have rodents).

Here are the five clades of mtDNA in FSL, also showing the genetic relationship of the other four subspecies with the inclusion if an outgroup species, Felis margarita, the adorable sand cat.

Here’s a sand cat just so you can see how cute they are. 

When the authors looked at the geographical distribution of these clades before domestication 10,000 years ago, they got the map below. Pay particular attention to the clades of FSL, as they are the ones used to trace the movement of cats in association with humans. A and B are in the Middle and Near East, and are genetically distinct from each other and from the DNA in clade C, which occurred (and still occurs) in North and Central Africa.

So looking at the genes of ancient cats from 10,000 years ago through Egyptian times and Roman times and medieval times to modern times, the authors found that the FSL subspecies appears to have been domesticated twice, It happened first in the Middle or Near East about 10,000 years ago, which we already knew, and then the descendants of those cats spread into southern Europe about 6000 years ago (we know this because cats from that region now carry the A and B mtDNA haplotypes seen above). Then there was another round of domestication that began in ancient Egypt about 4000 years ago. We know from ancient writings and artwork, as well as cat mummies, that the Egyptians kept tabby cats, and that their cats had the haplotype C from that region. Here’s an ancient painting from the Ottoni et al. paper showing an Egyptian tabby eating a fish under a woman’s chair. As the paper notes:

The image shows a ‘cat under the chair’ with a tabby mackerel marking, typical of F. silvestris lybica (Anna (Nina) Macpherson Davies, Copy of Wall Painting from Private Tomb 52 of Nakht, Thebes (I, 1, 99–102) Cat Eating Fish. Photo: © Ashmolean museum, Oxford, UK).

As shown by the correlation of clade C’s mtDNA with specimens from dated sites, the Egyptian-domesticated cats also moved into Europe, and in fact those Egyptian descendants became more numerous in Europe than did cats descended from the Middle and Near Eastern clades.  One of these Egyptian-like cats was also found at a Viking trading port, Ralswiek, on the Baltic sea, suggesting that cats were being moved around on Viking ships. Some of the Egyptian-clade domesticated cats even made their way to the region that’s now Iran, the domain of the wild subspecies Felis silvestris ornata.

So, based on mtDNA (and this needs confirmation with nuclear DNA, since mtDNA is really just a single gene that shows no recombination), cats in Europe are derived from FSL that was domesticated twice—in two places and thousands of years apart.

There’s one more interesting finding: the authors were able to get an idea about when humans began selectively breeding cats for coat patterns. (Cats probably largely underwent both natural and artificial selection for tameness, with the tamer wildcats able to get more food by approaching human settlements more closely, and then people breeding those cats that tended to hang around.)

There is one gene that is an indication of human selection for pattern: the gene causing the appearance of blotched rather than mackerel tabbies. As you see above, F. silvestris are mackerel tabbies in the wild, while the blotched pattern, seen below, is found only in domestic cats:

Blotched pattern on a tabby

We also know that the difference between striped (“mackerel”) and blotched tabbies resides at a single gene, Taqpep (a “transaminopeptidase”), with the blotched form being recessive to mackerel (you need two copies of the blotched gene to get the blotched pattern). We also know the precise DNA sequences that code for either the mackerel or blotched pattern. The authors were able to get nuclear DNA sequences of this gene from about 90 cats, and ten of these had the blotched form, with this gene appearing earliest about 1400 AD.  Conclusion? Probably that, at least for coat pattern, people didn’t select cats for a preferred appearance until medieval times.

This is a very cool study, and course is especially interesting to us ailurophiles. What we need now are more data using nuclear rather than mtDNA (the latter tends to move more readily between populations), and, especially, samples from the Far East and elsewhere in the world, since all the samples studied came only from Africa, Europe, and the Middle and Near East. What happened in China? Were cats domesticated there 5,000 years ago, as one study suggests, and, if so, were they from F. silvestris ornata or F. silvestris bieti, whose ranges extend into the Far East? Time will tell, my cat-loving friends, but be assured that, given the big public interest in felids, we’ll have the genetic data soon.


Ottoni, C., W. Van Neer, B. De Cupere, J. Daligault, S. Guimaraes, J. Peters, N. Spassov, M. E. Prendergast, N. Boivin, A. Morales-Muñiz, A. Bălăşescu, C. Becker, N. Benecke, A. Boroneant, H. Buitenhuis, J. Chahoud, A. Crowther, L. Llorente, N. Manaseryan, H. Monchot, V. Onar, M. Osypińska, O. Putelat, E. M. Quintana Morales, J. Studer, U. Wierer, R. Decorte, T. Grange, and E.-M. Geigl. 2017. The palaeogenetics of cat dispersal in the ancient world.  Nature Ecology & Evolution, doi:10.1038/s41559-017-0139

Supplementary information and data here.


  1. Posted June 21, 2017 at 10:11 am | Permalink

    Great stuff, thanks for your explanation of the findings in this paper.

  2. CJColucci
    Posted June 21, 2017 at 10:13 am | Permalink

    Until today, I had never realized the significance of naming the iconic Warner Brothers puddy tat “Sylvester.”

  3. Kevin
    Posted June 21, 2017 at 10:19 am | Permalink

    One of my cats can catch a mouse a day (and she does). The other one, barely a lizard. And yet the other never catches anything, but they wait for the other cat to catch it and then nom the leftovers.

    Why the difference? Makes me think cats can have the ability to partially rely on one another for survival or they are smart enough to recognize the advantages of their nearest neighbor.

    • Posted June 21, 2017 at 10:32 am | Permalink

      In the wild, I’m sure that F. silvestris always goes for mice. They may have lost some of their evolved hunting abilities under domestication, but, as we know, even placid housecats will hunt birds, mice, and insects (look at Hili!)

  4. Reginald Selkirk
    Posted June 21, 2017 at 10:28 am | Permalink

    I thought we were past the days when you could publish a meaningful study based only on mitochondrial DNA and one nuclear gene.

    … and ten of these had the blotched form, with this gene appearing earliest about 1400 AD. Conclusion? Probably that, at least for coat pattern, people didn’t select cats for a preferred appearance until medieval times.

    Of course a trait would not be selected until it appeared.

    • Posted June 21, 2017 at 10:34 am | Permalink

      There was a lot of work in that study, and I think the results, though they may be preliminary, justify publication,

      As for the gene needing to be there before it was selected (appearing via mutation), of course you’re right, but it doesn’t show up at all in early samples, and is more frequent after medieval times. That probably means that somewhere around 1400 a recessive homozygote appeared and was selected. The gene was certainly there, but rare, in earlier cats, and wasn’t found by the authors.

  5. Merilee
    Posted June 21, 2017 at 10:52 am | Permalink


  6. Jacques Hausser
    Posted June 21, 2017 at 11:06 am | Permalink

    Very interesting post. No time to read the paper in details just now, but I wonder: Is it possible that the non-contribution of the European F. sylvestris sylvestris to the domestic cat could come from the mere fact that it is, well, a forest cat, thus not attracted by the artifical open country cleared by agriculture? In contrast, F.s. lybica lives in open country and therefore was “pre-adapted” to agriculture.
    They are wild cats in my region, but they are very secretive and I have seen one of them in the wild only in three occasions. Impressive and mighty animals, looking at you with disdain before disappearing behind trunks and rocks.

    • W.Benson
      Posted June 21, 2017 at 12:30 pm | Permalink

      Good comment! Denning sites may be difficult for cats to find in deserts which might predispose native cats to den in junk or ruins around habitations. This might work for north Africa but perhaps not so much for Turkey.

  7. Posted June 21, 2017 at 11:08 am | Permalink

    This is great. Thanks.

  8. Torbjörn Larsson
    Posted June 21, 2017 at 11:14 am | Permalink

    So cats were domesticated at least 2 times? And humans did not select for cat appearance until lately.

    The BBC article has one researcher making the obvious conclusion:

    “There was very little breeding and selection going on in cats up the 19th Century, in contrast with dogs,” said Dr Geigl. “The cat was useful from the very beginning – it didn’t have to be changed.”

    [ ]

    Sadly, Dr. Geigl did not use the correct cat-technical term.


    • Posted June 21, 2017 at 12:00 pm | Permalink


    • Posted June 21, 2017 at 1:03 pm | Permalink

      or even purrrrrfeect. 🙂

    • rickflick
      Posted June 21, 2017 at 1:20 pm | Permalink

      It occurred to me, dogs are very different primarily in that they are social animals. That means they would have been easier to tame, but not very useful until they could be used for guarding the village or in hunting.

      • Posted June 21, 2017 at 3:09 pm | Permalink

        I suspect they were useful for food, as they still are in some cultures. Like horses.

        • gravelinspector-Aidan
          Posted June 21, 2017 at 5:48 pm | Permalink

          I suspect they were useful for food,

          The Dibbler criterion?

  9. Posted June 21, 2017 at 12:05 pm | Permalink

    I work in cancer biology and the scientific papers I read are interesting to me in a technical and professional sense, but rarely are they so evocative of things outside of molecular and cell biology. Papers like this are so much fun to read because of the images they evoke.

    The authors deduced that the Egyptian variant of kitties “… must have been very popular, as IV-C1 and C* represented more than half of the maternal lineages in Western Anatolia during the 1st millennium AD, and occurred twice as frequently as the local mitotype IV-A*.” This they speculate is due to “…success of the Egyptian cat is underlain by changes in its sociability and tameness.” It is easy to see today how this must have been true then. Ancient peoples loved their cats just as we do today -but couldn’t or wouldn’t change them like they did their d*gs. I like to think that our ancestors didn’t try to change cats because deep down they knew it would so change their nature and our relationship to them that it wasn’t the right thing to do.

    Anyway, the authors go on to trace the spread of moggies from port to port as neolithic and early modern humans spread, from highest Scandinavia (the Viking port of Ralswiek) to far into the interior of Eurasia (the port of Berenike on the Red Sea). They found evidence of cats moving with people along the old Silk Road in Turkey to south Asia and the Mediterranean basin via the Indian Ocean and Red Sea. They followed pussycats as far south as East Africa. It’s like a travelogue of early human society but done by tracing the steps of their companions, their familiars. It evokes a long, complex and interdependent relationship between two species during key developments in both their evolutionary histories. Fascinating.

    All this from looking at SNP variations in mtDNA from cat bones and dried up mummies. Thanks, WEIT.

    Sorry for the long windedness.

  10. Posted June 21, 2017 at 12:07 pm | Permalink

    Thank you for the helpful analysis of the study. My blotched tabby, Charlie (Darwin), has called my attention to this related post

  11. Mark R.
    Posted June 21, 2017 at 2:05 pm | Permalink

    Very intriguing…as you noted, it would be interesting to see how Chinese/Korean/Japanese cats fit into this picture. As well as SE Asian kittehs.

    • loren russell
      Posted June 21, 2017 at 10:28 pm | Permalink

      Other pubs I’ve seen suggest that present-day domestic cats are F.s.lybica all the way down. [excepting modern deliberate hybrids with other small felids]. The very early Chinese anthropolocal cats are outside the range of F.s.lybica so must have been another subspecies or even another species, but there doesn’t seem to be evidence that they have contributed to modern populations.

      The Siamese is clearly a divergent form, but it’s thought to be selected or drifted from F.s.lybica — helped by the fact that there are no F.sylvestris in se Asia, so no back-crossing to wild populations.

      Much or most of the divergent morphology and coat of Siamese and the allied “Oriental” breeds is the result of unnatural selection in the past 200 years. See debate about the aesthetics of “apple head” versus “wedge head” in the breed literature.

    • Posted June 22, 2017 at 11:54 am | Permalink

      I wondered this as well, especially with the recent posts about Japan.

  12. gravelinspector-Aidan
    Posted June 21, 2017 at 5:57 pm | Permalink

    the geographical distribution of these clades before domestication 10,000 years ago, they got the map below. Pay particular attention […] clade C, which occurred (and still occurs) in North and Central Africa.

    That cluster of clade C around Lakes Victoria and Tanganyika really sticks out like a tail near a rocking chair. At 10kyr BP!
    The obvious suggestion is that a group of Egyptians (and accompanying cats) made their way up the length of the Nile before 10kyr BP. Which is not impossible. But is challenging.
    I’ll read the paper.

    • gravelinspector-Aidan
      Posted June 21, 2017 at 6:25 pm | Permalink

      The caption (SuppInfo) is that the clade regions are inferred pre-historic ones, but the table of genotypes gives the entries for East Africa (location groups 27-31) as being post-1900 CE (27, 28, 29, 31) and 500-1300 CE (30). Which I’d infer as being Arab trading (pre-about 1850) followed by European colonisation. Essentially, ship-born(e) cats, not pre-historic feline precursors to Burton and Speke.
      Not that I’d be particularly surprised to find an Egyptian expedition up the Nile at (say) 3kyr BP (1000 BCE) – they went down into Sudan around that time. But at 10kyr BP … eyebrows raised.

  13. Posted June 21, 2017 at 11:30 pm | Permalink

    Reblogged this on The Logical Place.

  14. Posted June 22, 2017 at 2:11 am | Permalink


  15. Posted June 22, 2017 at 2:12 am | Permalink


  16. Hempenstein
    Posted June 22, 2017 at 8:33 am | Permalink

    Vikings travelled and traded widely, ranging down into the Mediterranean, so they could easily have been catapulted from Egypt in one jump.

One Trackback/Pingback

  1. […] keep in mind that, comprehensive as it was, this was a study of cat mtDNA alone–pretty much a study of a single gene that doesn’t recombine, according to Jerry Coyne at Why […]

%d bloggers like this: