The complex evolution of the big cats

A new paper in Science Advances by Henrique V. Figueiró et al. (reference below, free download with legal UnPaywall app) resolves the contested phylogeny (family tree) of five big cat species: lion, leopard, jaguar, snow leopard, and tiger. In so doing, the researchers uncovered some interesting evolutionary history.

Previously, the family tree of this group had been a bit confusing, as different genes gave different phylogenies. That’s normal for recently evolved species (the speciation in this group occurred within the last 4.5 million years), for the genes that were variable in the common ancestors could remain variable in the descendants, and depending on which gene you looked at, you could find evolutionary phylogenies that were really the histories of the genes themselves rather than the history of the species that contained them. But there was another reason for the confusing phylogenies as well, and we’ll get to that shortly.

I’ll try to be brief so we don’t have too many long posts today. Figueiró et al. did whole-genome sequencing of these species, which is not so hard these days, and based on the consensus relationships of all the genes, they came up with this phylogeny. The numbers above the lines of splitting give the estimated times at which those splits occurred:

The African cats are the most closely related, with the jaguar (a south American cat) splitting off from the ancestor of the leopard and lion about 3.6 million years ago. But how can that be, since Africa and South America themselves split over 150 million years ago? Yes, you guessed it: the jaguar evolved in Eurasia, and moved down to South America after crossing to North America. Jaguars once lived in what is now England!

The Asian cats—the tiger and snow leopard—are also “sister species”—each other’s closest relatives—having split about 3.5 million years ago. (The tree was “rooted” by using a species that wasn’t in the group of large cats: somebody’s moggie from Missouri.)

By looking at how fast DNA sequences evolved compared to the rest of the genome, each branch of the Big Cat Tree shows which class of genes show members subject to “positive selection”—genes favored by especially strong natural selection. Jaguars and tigers show evolution, for instance, of genes known to be involved in protein metabolism and smell; and jaguars, which have big thick heads compared to the other cats, have positive selection on genes involved in “craniofacial development.”

By looking at gene sequences and comparing them with population-genetic models, you can also get an idea of the number of individuals in the species at various times in its history (well, an underestimate of the size). Here’s that history for the big cats, and it shows two episodes of population reduction: one about 100,000-300,000 years ago, and a second round between 10,000 and 20,000 years ago, possibly connected with glaciation. Historically, jaguars, despite their large former range, seem to have had the lowest population size averaged over time.

One of the study’s striking findings was the degree of genetic interchange between branches of the tree that had already split; that is, between species. Like humans and Neanderthals, big cat species occasionally exchanged genes after the species had gone their separate evolutionary ways. This plot shows which branches exchanged genes as the ancestors gave rise to their descendants:

It is in fact this exchange of genes between ancient branches (not necessarily, for instance between living lions and jaguars but between the ancestors of each) that probably made the phylogenies confusing. If you looked just at the genes exchanged not that long ago between the ancestors of modern lions and jaguars, you’d come to a wrong conclusion: that lions are more closely related to jaguars than to leopards. That’s why it’s important to look at many genes instead of just a couple when making these trees.

Finally, there’s a striking association between the genes exchanged between species and the genes that have evolved quickly by natural selection. Genes that were exchanged between lineages and then evolved rapidly after those exchanges include genes affecting brain function and development, genes involved in the growth and guidance of axons (nerve connections), and genes affecting the optic nerves. The authors suggest that this means that “introgression” (genes coming via hybridization with another species) is a useful source of adaptive variation: a reservoir of variation that natural selection can use. (It’s analogous to mutation, which also introduces new variation into the genome.)

But I have another interpretation: a gene that gets in from another species one way or the other could be subject to rapid selection because it’s suddenly found itself in a foreign genome to which it’s not adapted, and has to change rapidly to “fit in” with the new genome rather than adapting to environmental circumstances. Now the first interpretation has clearly happened, as we see in introgressed butterfly color and pattern genes that are used, after hybridization, to make new mimetic patterns; but I don’t see that my interpretation can never apply.

Regardless, we now know the evolutionary history of this group of large cats (they’re more closely related to each other than to any other species of felid), and we’ve learned that genetic interchange between separated branches of the evolutionary tree is more extensive, in this and other groups, than we ever expected.

h/t: Matthew Cobb


Figueiró, H. V., et al. (2017).  Genome-wide signatures of complex introgression and adaptive evolution in the big catsScience Advances 3(7). e1700299 DOI: 10.1126/sciadv.1700299


  1. rickflick
    Posted July 30, 2017 at 1:39 pm | Permalink


  2. ThyroidPlanet
    Posted July 30, 2017 at 2:13 pm | Permalink


  3. Liz
    Posted July 30, 2017 at 2:16 pm | Permalink

    “The authors suggest that this means that ‘introgression’ (genes coming via hybridization with another species) is a useful source of adaptive variation: a reservoir of variation that natural selection can use.”

    Is there any way then, that the hybridization of a horse and donkey is “useful” in this way? The offspring, mules, are barely fertile unless it’s a female mule mating with a male horse or donkey. At least that is my understanding. A male and female mule aren’t able to reproduce.

  4. John Aylwin
    Posted July 30, 2017 at 2:18 pm | Permalink


    Could it not also be argued that the “rapid selection” applies not just to the infiltrating gene, but also to the existing genes in the genome in which it finds itself? The selection pressure is on all genes in their context, and that context includes the existing genome (as well as the environment). The existing genes now have a variation of context to be played with.

  5. Michael Fisher
    Posted July 30, 2017 at 2:28 pm | Permalink

    Very nice! It’s a pity the Missourian moggie isn’t listed by her handle. I was bit suspicious of the name of one of the thirty-odd authors, a Leandro Silveira [Leo & silvestris]! Could he be a furry author? After all, you can’t have a more catty name without invoking “Felix” or “Meow Two Tongue”

    Unfortunately he’s a real human catologist & president of the Jaguar Conservation Fund. So the moggie goes unnamed.

  6. Michael Fisher
    Posted July 30, 2017 at 2:52 pm | Permalink

    Genuine question. Not snark.

    “…the degree of genetic interchange between branches of the tree that had already split…” and “we’ve learned that genetic interchange between separated branches of the evolutionary tree is more extensive, in this and other groups, than we ever expected”

    This puzzles me – for it’s only a period of say 10 million years for these six species to sprout from one species. The various splits are fuzzy as expected & it’s a short time, so how can ‘fuzziness’ be described as “more extensive […] than we ever expected”? Is there an expected amount of ‘fuzziness’ per million years of speciation?

    • Posted July 30, 2017 at 4:40 pm | Permalink

      What I meant is that until recent years, the idea was that there was virtually no genetic interchange between “good species” because of reproductive isolation. We know now that while introgression isn’t rampant, it can occur when speciation is well along but not complete, or even between “good” species that are not 100% reproductively isolated.

      • Perluigi Ballabeni
        Posted July 31, 2017 at 6:43 am | Permalink

        This suggests we need a second edition of your speciation book.

    • Torbjörn Larsson
      Posted July 30, 2017 at 5:52 pm | Permalink

      My interpretation is that the interchange is unusually large, likely mostly because we are still filling in large lacunae of sequencing. (But also because (large) cats are known for this, if I remember Jerry’s earlier articles correctly.)

      • Torbjörn Larsson
        Posted July 30, 2017 at 5:59 pm | Permalink

        Sigh, forgot to update again. But I was not too far off, I think. (The earlier articles were describing large cat crossings between zoo animals, if I remember correctly.)

  7. Alan
    Posted July 30, 2017 at 3:49 pm | Permalink

    Interesting. The scenario you describe could describe the purging of polymorphic incompatibilities linked to the selected variant. An introgressed variant under selection may be linked to other variation (that evolved in the other species) associated with incompatibilities. This gene sequence could then evolve for genomic compatibility as the polymorphic DMIS are purged – “to ‘fit’ in” – in the new genomic environment.

  8. Dave
    Posted July 30, 2017 at 4:35 pm | Permalink

    “The African cats are the most closely related, with the jaguar (a south American cat) splitting off from the ancestor of the tiger and lion about 3.6 million years ago.”

    From looking at the tree, I think this should say “leopard and lion”, not “tiger and lion”. The tiger/snow leopard clade is the sister-group to the lion/leopard/jaguar clade.

  9. Mark R.
    Posted July 30, 2017 at 5:03 pm | Permalink

    Did you write about “positive selection” in WEIT? I don’t remember you mentioning it if you did. Thanks for this lucid explanation of an evolutionary mechanism I didn’t previously know.

    Is there sabretooth tiger DNA preserved well enough to map its genome; and if so, is it mapped? I would be interested to know where sabretooths fit in this tree.

    • Posted July 30, 2017 at 6:54 pm | Permalink

      Current ideas are that the true sabertooths split off before the divergence of the living felids. And yes, they are just recently getting good DNA from sabertooth fossils.

      • Mark R.
        Posted July 30, 2017 at 7:03 pm | Permalink

        Thanks for the information Mr. Felenstein, I appreciate it.

  10. Posted July 30, 2017 at 7:06 pm | Permalink

    Thank you for highlighting this paper.

  11. Posted July 30, 2017 at 7:30 pm | Permalink

    Very interesting! This would be a good lesson in interpreting phylogenies.

  12. Rob Munguia
    Posted July 30, 2017 at 7:34 pm | Permalink

    Thanks for the nice science post PCC.

  13. johnw
    Posted July 30, 2017 at 8:22 pm | Permalink

    This complex divergence with subsequent admixture reminds me of modern humans:Neanderthals:Denisovans:possibly H. Erectus admixture which has been recently described in ancient DNA work. Also, the present day story of the gray wolf :coyote:Eastern wolf:red wolf story in the US, which has been described as a hybrid swarm where the species overlap, though certain combinations only seem to occur if humans intervene (western gray or timber wolf : coyote).

    • Diane G.
      Posted July 31, 2017 at 3:28 am | Permalink

      I love the current North American canid story–real-time evolution in action, in a clade of relatively large carnivores no less. 🙂 IIANM, though, it’s threatening conservation efforts for the red wolf, at least in the Carolinas.

  14. Posted July 30, 2017 at 8:57 pm | Permalink

    Thanks for this science post.

  15. Posted July 31, 2017 at 2:27 am | Permalink

    Reblogged this on jtveg's Blog and commented:
    Big cat evolution!

  16. Alan Clark
    Posted July 31, 2017 at 3:42 am | Permalink

    i am surprised that the tree shows Lions and Tigers so distantly related, when they can interbreed in captivity. What is the explanation for this?

    • Reginald Selkirk
      Posted July 31, 2017 at 7:22 am | Permalink

      Camels and Llamas can also interbreed, with human assistance.

    • darrelle
      Posted July 31, 2017 at 8:02 am | Permalink

      Perhaps relatively speaking, relative to just the small group of big cats, they are distantly related. But one thing this study shows is that generally speaking all of the big cats are rather closely related. Per this study the last common ancestor of all of the big cats was only 3.37 – 5.84 million years ago.

    • Diane G.
      Posted August 1, 2017 at 2:03 am | Permalink

      IIANM, sometimes geographical isolation serves as enough of a reproduction barrier that there’s no particular need for selection of genetic barriers…

  17. Reginald Selkirk
    Posted July 31, 2017 at 7:26 am | Permalink

    Pumas (mountain lions) got left out? If jaguars got to South America by passing through North America, then pumas and jaguars must have shared a range for a while.

    • darrelle
      Posted July 31, 2017 at 8:16 am | Permalink

      I think they still might. There are mountain lions in South America, though I don’t know off hand if their ranges overlap anywhere with jaguars.

      Taking a quick google it looks like their ranges do overlap and that where they do the puma is usually smaller than average.

      Also, apparently cougar are typically not grouped with the big cats and are thought to be more closely related to smaller cats than to the big cats. I think it would have been better if this study had included cougars so that the cougar’s relationship among cats could have been better understood.

      • Martin Levin
        Posted July 31, 2017 at 1:30 pm | Permalink

        The average cougar is larger and heavier than either leopards or snow leopards, though streamlined, with a smaller skull and neck, and is decidedly a big cat.

        • darrelle
          Posted July 31, 2017 at 2:23 pm | Permalink

          That may be so, but for reasons other than physical size cougar’s are not typically grouped with the “Big Cats.” They are typically considered to be the largest of the “Small Cats.” That’s not an opinion on my part, merely an observation of what the experts say. For example, here is an excerpt from the Taxonomy section of the Wikipedia entry on Cougar.

          “Cougars are the largest of the small cats. They are placed in the subfamily Felinae, although their bulk characteristics are similar to those of the big cats in the subfamily Pantherinae.[1] The family Felidae is believed to have originated in Asia about 11 million years ago.

          And from the Physical Characteristics section.

          “The cougar is on average larger than all felids apart from lions, tigers, and jaguars. Despite its size, it is not typically classified among the “big cats”, as it cannot roar, lacking the specialized larynx and hyoid apparatus of Panthera.[49]

          Given the several surprises discovered by this new whole genome study I wonder what may have been found regarding the cougar’s relationship to the “Big Cats” if they had included the cougar.

  18. Ken Elliott
    Posted July 31, 2017 at 8:04 am | Permalink

    How do cheetahs relate to these big cat species? I am a total layman, so please forgive me if the answer is obvious to more knowledgeable folks.

  19. nicky
    Posted July 31, 2017 at 2:08 pm | Permalink

    Since lions and tigers do hybridise, and are furthest from each other on the lineage, can we now surmise that, say, lions and leopards probably could too? Can all big cats hybridise with each other? Has that ever been attempted?

  20. Avis James
    Posted July 31, 2017 at 9:16 pm | Permalink

    Thanks for more information for my classes!

Post a Comment

Required fields are marked *

%d bloggers like this: