A new study of polar bears underlines the dangers of reconstructing evolution using mitochondrial DNA

Biology Lesson:  DO NOT MAKE EVOLUTIONARY TREES OF ANIMALS AND PLANTS BASED ENTIRELY ON MITOCHONDRIAL DNA (mtDNA): PLEASE USE NUCLEAR DNA WHENEVER YOU CAN.  THIS IS BECAUSE mtDNA APPEARS TO MOVE MORE READILY BETWEEN SPECIES THAN DOES NUCLEAR DNA (nDNA), CAUSING A DISCORDANCE BETWEEN EVOLUTIONARY TREES BASED ON MITOCHONDRIAL GENES (‘GENE TREES’) AND THOSE BASED ON POPULATION AND SPECIES HISTORY THAT ARE DISCERNED FROM ANALYSES OF MANY NUCLEAR GENES (‘SPECIES TREES’).

I have put that in capslock because I’ve been emphasizing this problem for a long time, and yet many systematists and evolutionary geneticists still persist in using the DNA from mitochondria (the energy-producing organelles in cells that stem from ancient bacteria and have their own DNA) to make evolutionary trees of organisms and estimate their divergence times.  They do this because 1) mtDNA is much easier to purify and sequence than is DNA from the nucleus (“nuclear DNA”, or nDNA), and 2) mtDNA usually evolves much faster than nDNA, supposedly making mtDNA a better indicator of species relationships, since it sorts out into definitive patterns more rapidly.

The problem is that, for reasons we don’t fully understand, mtDNA also moves between species during hybridization much more readily than does nDNA, and that can screw up species relationships. This is true for both animals and plants, and not just for mtDNA either: in plants, DNA from another organelle, the chloroplasts (site of photosynthesis, this DNA is called “cpDNA”) also moves between species more readily than nDNA.

Although the species definition used by most evolutionists, the “biological species concept” (BSC) uses the presence reproductive isolation between groups (mate discrimination, ecological preferences, hybrid sterility) as the criterion for their status as separate species, those reproductive barriers aren’t always complete, and sometimes genes can leak between species via hybridization. (The hybrids have to be partially fertile to transfer genes between species).

In the pair of Drosophila species I’ve worked on for 15 years, for example, the mitochondria from one species have completely replaced those of another species on the island of São Tomé.  If you made a phylogeny of those species based entirely on mtDNA, you’d find that on the island they appear to have diverged only a short time ago, and that they did so on the island (“sympatrically”). But analysis of nuclear DNA shows that this is wrong: the two species are about 400,000 years old, and diverged when one ancestor invaded the island, formed a new species, and then a second wave of invasion from the second ancestral group re-invaded, yielding the two sister species on the same island.

We see this situation over and over again in biology. We don’t really know why mtDNA (and cpDNA) leak so readily between species, but we do know that this leakage makes it dicey to use only organelle’s DNA to make species trees.  But the reason for this leakage compared to nDNA (so common to be almost a “rule of biology”) would make a useful paper topic for some enterprising graduate student.

But enough harangue.  The misleading evolutionary conclusions one can draw from using mtDNA alone are well demonstrated in a new study of the phylogeny of polar bears and brown bears published by Miller et al. in Proc. Nat. Acad. Sci. (reference below; free access and download; see also J. D. Gorman’s summary in today’s New York Times).

The phylogenetic position and age of the polar bear (Ursus maritiumus) versus its closest relative, the brown bear (Ursus arctos), has for some reason become a lively topic of research. A while back I posted (here) that a population of brown bears on Alaska’s Alexander Archipelago (called “ABC brown bears”) might be more closely related to polar bears than to other brown bears.  If that were true it would mean that the species “brown bear” was “paraphyletic,” i.e., some populations of brown bears are more closely related to another species than to other brown bears. For some biologists, that menas that brown bears wouldn’t be qualified as a “species” .

In a paper by Lindqvist et al. in 2010 (ref. below), the divergence time between polar and brown bears was estimated at about 150,000 years: a remarkably short time for a speciation event.  A more recent study published this year in Science (reference also below) used nDNA to estimate a divergence time that is more reasonable: 600,000 years.  But they used only a limited sample of nuclear DNA.

In my post on the Lindqvist study, I cautioned people about trying to reconstruct population history from mtDNA alone because its easy movement between species makes it a bad candidate for reconstructing the history of populations and species.  That warning now appears reasonable based on a full-genome analysis of polar and brown bears (as well as black bears) reported in the PNAS paper  noted above.

Now that we can sequence full genomes fairly rapidly and cheaply—this amazing advance has occurred in only the past 30 years—it’s possible to reconstruct evolutionary histories using nearly every gene in an animal or plant, and that’s as close as we can get to an accurate reconstruction without having been around when evolutionary lineages diverged.  And that’s what Miller et al. did: they sequenced genomes of one polar bear, two ABC brown bears, a non-ABC brown bear, and a more distant relative, the American black bear (Ursus americanus). As they state in their paper, the goals were these:

We gathered extensive genome sequence data from modern polar, brown, and American black bear samples, plus a ∼120,000-y-old PB, to address the following questions. (i) What is the more precise association between the PB and its sister species, the brown bear; and do we find any signatures of past genetic interchange between the two species? (ii)Did the PB indeed evolve recently, as suggested by mitochondrial DNA and fossil evidence, or did it have an older origin, as demonstrated by nuclear DNA loci? (iii) Can we deduce any past responses in ancient bear population histories that may be connected with climatic changes?

Here are the results, summarized briefly:

  • Polar bears diverged from brown bears about 4-5 million years ago, shortly after (i.e., within a million years) their common ancestor diverged from the ancestor of the black bear, the next most closely related species.  Polar bears, then, are far older than previous studies suggested.
  • The young age previously estimated from mitochondrial DNA suggests that that DNA was moved between polar and brown bear ancestors by hybridization after the lineages had been diverging.  This is not unreasonable, since polar and brown bears are able to hybridize and produce fertile offspring in zoos, and appear to do so rarely in the wild.  This suggests that the speciation event producing modern polar and brown bears was sporadically interrupted by hybridization and gene flow between them, probably because climate change forced them to encounter each other when their ranges moved.

The figure below shows the discordance between family trees based on mtDNA and nuclear DNA: the tree based on mtDNA is in orange, that based on many nuclear genes is given by the white bars outlined in black. Notice the deceptively recent divergence time between brown and polar bears indicated by mtDNA, and that suggestion that in one instance the polar bear was more closely related to a brown bear than the tw0 brown bears are to each other. This “paraphyly,” too, is an illusory result.

  • As the authors note, “The clear discordance between mitochondrial and nuclear genomes in the phylogenetic placement of the ABC brown bears mirrors that found in the evolutionary histories of archaic and anatomically modern human lineages.” What they mean is that up to several percent of modern human DNA comes from Neanderthals, probably via ancient hybridization. If one looked only at that DNA, one would reach the false conclusion that some modern humans are more closely related to Neanderthals than to other modern humans.  This is the danger of using only a limited amount of DNA to draw conclusions about evolutionary history.  Some ABC brown bears share as much as 11% of their genes with polar bears!
  • The 4-5 million-year-old split between black bears and the polar/brown bear lineage was also followed by hybridization lasting until about 100,000 years ago.  In contrast, the near-simultaneous split between the sister species of brown and polar bears was followed by hybridization that can’t be said to have stopped even today.  Genes are still, to our knowledge, occasionally exchanged between these species.  In that sense we can’t say they are “complete” biological species, but are “nearly complete” biological species. Such is the case when there is sporadic gene exchange between groups whose reproductive isolation is nearly but not fully impeded by biological barriers.
  • Based on genetic reconstructions, the polar bear population has declined drastically during the last half million years; the authors impute this to (nonanthropogenic) warming of the climate.

The overall lessons are these.  First, use nuclear DNA to reconstruct evolution whenever you can, and always be suspicious of evolutionary trees based solely on mitochondial or chloroplast DNA. Second, polar bears and brown bears diverged a long time ago, not, as other studies suggested, quite recently.  Third, the polar/brown bear divergence, as well as their joint divergence from black bears, involved occasional hybridization, so that speciation in these cases did not involve the classic scenario of complete geographic isolation up to the point where genomes diverged to the point of immiscibility. As Gorman says in his NYT piece: “The progress of species formation, at least in this case, is a bit like a long, ambivalent divorce in which the two parties separate but occasionally fall back into bed even after the official decree.”

Finally, polar bears are sensitive to climate, and as we continue to heat up our environment through shortsightedness, those bears are liable to extinction. They will either be unable to support themselves ecologically as the polar ice disappears, or they’ll hybridize themselves out of extinction by mating with brown bears. Either way, the fate of this lovely animal is precarious.

Humans poison everything, and I see no way, given our greed and dependence on fossil fuels, that we’ll be able to stop the trend of global warming.

Whatever, just remember to conserve energy and be wary of mitochondrial DNA.

Don’t extinct me, bro!

_________

Hailer, F. et al. 2012.  Nuclear genomic sequences reveal that polar bears are an old and distinct bear lineage.  Science 336:344-347.

Lindqvist, C. et al. 2010.  Complete mitochondrial genome of a Pleistocene jawbone unveils the origin of polar bear. Proc. Nat. Acad. Sci. USA 107:5053-5057,

Miller, W. et al. 2012.  Polar and brown bear genomes reveal ancient admiture and demographic footprints of past climate change. Proc. Nat. Acad. Sci. USA, Published online before print July 23, 2012, doi: 10.1073/pnas.1210506109

95 Comments

  1. gbjames
    Posted July 24, 2012 at 5:37 am | Permalink

    What is that little dead-end polar bear branch?

    • Posted July 24, 2012 at 6:30 am | Permalink

      According to the caption in the original paper, it represents the time at which hybridization caused replacement of the original polar bear mitochondrial genome with the ABC brown bear’s.

      • gbjames
        Posted July 24, 2012 at 6:44 am | Permalink

        Thanks. I missed that.

      • Torbjörn Larsson, OM
        Posted July 24, 2012 at 1:01 pm | Permalink

        Wouldn’t that replacement be the cross gene extinction, and the branch be an actual diverged PB population?

        • Torbjörn Larsson, OM
          Posted July 24, 2012 at 1:34 pm | Permalink

          No, apparently fig 3 shows two different hypotheses of lineage splitting & admixture. The dead end is the ancient PB in one of them. Supposedly the cross is the other, but the figure doesn’t really say. (It is badly referenced, and claim to use the grays for illustration of both hypotheses.)

          • Posted July 24, 2012 at 5:23 pm | Permalink

            Yeah I think I misunderstood what gbjames was asking about – the dead end with the “X” in the mtDNA tree is the hybridization event; the dead end in the species tree is the 130,000 year old polar bear specimen from Svalbard as Torbjorn points out.

            (And that grey shading! How on earth are we supposed to understand that?!)

  2. Posted July 24, 2012 at 5:45 am | Permalink

    Can we rule out chance or accident as the explanation for why some invasive mitochodria become fixed in some populations? Is the frequency really higher that for any nuclear gene with distinctive alleles?

    • AnatomyProf
      Posted July 24, 2012 at 6:18 am | Permalink

      I was wondering that. It seems that it would be possible for a female to bring in new mitochondria and have it over represented in the population as a result of producing female offspring. does it simply suggest that in the average hybridization event involves a female of the less represented species or species where the offspring cross with the male’s species?
      I would think that you could tackle this question with flies, but this isn’t something that I know anything about. I don’t even remember discussion of this topic in evolution or conservation bio classes. I would however like to be able to mention to students any hypotheses that might exist for this difference if someone is aware of them.

    • CJ
      Posted July 24, 2012 at 7:15 am | Permalink

      Hybridization isn’t really chance is it? If the mitochondria become fixed, then they are to some degree adaptable to their new, yet related host, and vice versa.

      i’d like to know the answer to your second question though.

      • CJ
        Posted July 24, 2012 at 8:39 am | Permalink

        i should have said, “to some degree adaptable and/or neutral.”

    • IdoP
      Posted July 24, 2012 at 7:40 am | Permalink

      I think we can’t rule out fixation by drift.

      For haploids the expected waiting time until fixation for a neutral allele is 2*Ne generations, where Ne is the effective population size. Looking at the appendix of the PNAS paper (fit S15), it appears that on average Ne has been about 10000. Assuming a generation time of 10 years, that gives a ball park estimated time to fixation of 0.2my, pretty close to the estimated 0.16my.

      • The Other Jim
        Posted July 24, 2012 at 1:08 pm | Permalink

        I think this is a bit simplistic, no?

        • Posted July 24, 2012 at 1:53 pm | Permalink

          Why is it simplistic?

          Random genetic drift is the null hypothesis. If the data is consistent with drift then why in the world would you postulate anything else?

          • The Other Jim
            Posted July 24, 2012 at 2:01 pm | Permalink

            Groan… I need to sleep…

            I read “we can rule out” not “we can’t rule out”. My apologies!

            -The Other Jim

        • IdoP
          Posted July 24, 2012 at 2:58 pm | Permalink

          Well, it may seem simplistic, but the underlying mathematics of diffusion equations is not that easy. I think it’s very cool that a simple model of neutral evolution can predict the observations so closely. My conclusion is that there is no need at all to invoke selection to explain the fixation of the mitochondria. Until new evidence suggests otherwise of course.

          • The Other Jim
            Posted July 25, 2012 at 12:22 am | Permalink

            As I commented above, I mis-read your 1st line as “can reject”, so I got your whole argument backwards.

            I promise to not post just before bedtime again 😉

  3. Bonzodog
    Posted July 24, 2012 at 5:46 am | Permalink

    Ignoring the science (which is excellent) what a wonderful piece ..

  4. Posted July 24, 2012 at 6:07 am | Permalink

    This is a great post, thanks very much for the scientific details. My co-workers and I are trying to figure out the evolutionary history of a group of orchids, based on a mix of genes. This post and the cited articles will make me much more cautious about our results.

    • gillt
      Posted July 24, 2012 at 2:16 pm | Permalink

      As JAC noted, everyone should be doing multi-locus analysis at the least. But why not use entire genomes? That’s the way the techbnology is moving. If you don’t have a reference genome to work from it might be worthwhile to dedicate one lane to whole genome shotgun sequencing for one of your speciemens then multiplex the rest.

      • The Other Jim
        Posted July 25, 2012 at 1:02 am | Permalink

        But why not use entire genomes?

        Money. One day, maybe, but for now…

        • gillt
          Posted July 26, 2012 at 4:49 pm | Permalink

          a gazillion reads for under $1k is pretty affordable. The real potential hurdle is what to do with all that data.

  5. Posted July 24, 2012 at 6:12 am | Permalink

    Simple question from a non-biologist: since, in mammals, mitochondrial DNA is only contributed by the ovum, why does it not give a reliable matrilineal phylogeny?

    Or is the whole point that where there is ocasional interbreeding, the matrilineal and true average “species” phylogenies can differ?

    • gbjames
      Posted July 24, 2012 at 6:19 am | Permalink

      I was wondering the same thing but gather that the answer is not yet known, given Jerry’s phrasing: “… for reasons we don’t fully understand…”. Apparently there is more to it than hybridization alone.

      • Posted July 24, 2012 at 6:33 am | Permalink

        It does, technically, give a reliable matrilineal phylogeny. It’s just that the matrilineal phylogeny conflicts with what might be considered the true species phylogeny. So it’s as Paul said, the point is that these can differ—and therefore if you only know the mitochondrial phylogeny, you might not know the full story.

    • Thanny
      Posted July 24, 2012 at 11:50 am | Permalink

      It takes a single hybrid mating to introduce an entire mitochondrial genome.

      Nuclear genes, however, have a 50% chance of not making it into the next generation at each mating. So if a female brown bear mates with a male polar bear, and that female line survives going forward (the “Mitochondrial Eve” of polar bears – kind of shows what a horrible name that is), all polar bears would have brown bear mtDNA dating back to that event. But absent more hybridization events, those brown bear nuclear genes would be diluted over time, potentially down to nothing. The end state could be polar bears all having brown bear mtDNA, but zero brown bear nuclear genes.

      As it turns out, that’s not exactly what happened, because some brown bears have 11% nuclear genes in common with polar bears, according to the study. But that still means that you have a line of polar bears down through time, with some brown bears thrown in the mix sporadically. If you want to know when that line separated from the brown bears for normal breeding purposes, you have to track the nuclear DNA. The mitochondrial DNA is a red herring, since it can be replaced entirely by just one female brown bear wandering near a horny male polar bear.

      • gbjames
        Posted July 24, 2012 at 11:56 am | Permalink

        This would be true for Y chromosome genes, too, right? (assuming bears use the same XY=male, XX=female system we do)

      • gbjames
        Posted July 24, 2012 at 11:56 am | Permalink

        This would be true for Y chromosome genes, too, right? (assuming bears use the same XY=male, XX=female system we do)

        • gbjames
          Posted July 24, 2012 at 11:56 am | Permalink

          Why two? Again? What’s going on? (Sorry for the duplicate posting.)

          • Posted July 24, 2012 at 1:08 pm | Permalink

            Diploid.

            /@

      • George
        Posted July 25, 2012 at 8:51 pm | Permalink

        So have they really sampled enough polar bears to say that every one alive today has mDNA from a hybridization event with a female brown bear 150,000 years ago? Or did they just sample 1 polar bear with this mDNA? If it was the case that all polar bears today have descended from that event, that would be really incredible.

        • Thanny
          Posted July 26, 2012 at 11:14 am | Permalink

          Quite improbable, I agree. It’s far more likely that some local population shared mtDNA from such an event, rather than the entire species.

          But I don’t know what the facts are.

        • Gregory Kusnick
          Posted July 26, 2012 at 11:32 am | Permalink

          Why would that be incredible? We know that all humans get their mtDNA from a common matrilineal ancestor about 200,000 years ago.

          The claim is not that hybridization happened just once, and that hybrid turned out to be the common matrilineal ancestor of all polar bears. (That would be a remarkable coincidence.)

          Rather, the assumption is that hybridization happened repeatedly, and the common matrilineal ancestor (which logically must exist) turns out to be one of the hybrids.

  6. eric
    Posted July 24, 2012 at 6:23 am | Permalink

    The phylogenetic position and age of the polar bear (Ursus maritiumus) versus its closest relative, the brown bear (Ursus arctos) has for some reason become a lively topic of research.

    I’m guessing there’s political reasons on both sides of the aisle. Business/conservative interests want to claim there is no reason to set aside habitat or reduce business in the area, because PBs are just a variant of brown bear. Conservationists want to show the opposite – but in addition, want to know how difficult it might be to reproduce PBs from brown bears should the worst happen.

    • gravelinspector
      Posted July 25, 2012 at 4:18 am | Permalink

      There may be “political” reasons. But there are also un-related contingent reasons. In this case, the chance discovery of a well-preserved polar bear skull several years ago in Scotland may be significant. (I downloaded the paper last night, by coincidence,but I haven’t got round to reading it.) That discovery was contingent on a lot of chance things, including the results of about 20 years of extremely severe cave-diving exploration, which lead to the discovery of the polar bear skull. (Acquaintances in the Grampian Speleological Group described some of the diving to me when I was playing that game ; it sounds spectacularly uninviting diving. By the standards of cave diving. In zero-visibility “fluid”. Snow-melt. In a ripped wetsuit. Six sumps. Each way. Enjoy!)
      I’m off to see the lawyers (yeuch!), and I’ll read the paper on the bus. This work may not be from the same fossil material, but there aren’t a lot of fossil polar bear skulls that are collected in DNA-friendly conditions. There has been a recent skull. There is a lot of recent polar bear DNA work. Coincidence?

      • Posted July 25, 2012 at 5:20 pm | Permalink

        can you cite this paper and/or say if they dated the skull?

        • gravelinspector
          Posted July 26, 2012 at 2:04 am | Permalink

          The discovery report is at http://www.sat.dundee.ac.uk/~arb/gsg/news/139.html
          Looking in the formal press, I don’t actually see any mention of it, though in chat at the Craven it has been mentioned as having gone to Glasgow for “further work.”
          With the current (relative) flurry of publications about polar bears, it looks as if I’d conflated the events, and there is no formal connection. So I’d just have to cite it as “Pers.Comm.”
          Sorry, my mistake. Having known some of the divers and cavers involved for decades, I’d kept that knowledge rattling in the back of the head, but forgotten that most people don’t know about such activities.
          OTOH … it’s something to look forward to, when it finally does get published.
          I’m trying to establish what the state of play is for this specimen.

        • gravelinspector
          Posted July 26, 2012 at 2:04 am | Permalink

          The discovery report is at http://www.sat.dundee.ac.uk/~arb/gsg/news/139.html
          Looking in the formal press, I don’t actually see any mention of it, though in chat at the Craven it has been mentioned as having gone to Glasgow for “further work.”
          With the current (relative) flurry of publications about polar bears, it looks as if I’d conflated the events, and there is no formal connection. So I’d just have to cite it as “Pers.Comm.”
          Sorry, my mistake. Having known some of the divers and cavers involved for decades, I’d kept that knowledge rattling in the back of the head, but forgotten that most people don’t know about such activities.
          OTOH … it’s something to look forward to, when it finally does get published.
          I’m trying to establish what the state of play is for this specimen.

  7. Reg Le Sueur
    Posted July 24, 2012 at 6:25 am | Permalink

    So what about “mitochondrial Eve”?
    I suppose that creationists might claim that her age of 200,000 years, must be wrong and nDNA studies makes her 6000 years old, –because the Bible tells us so.

    • john
      Posted July 24, 2012 at 8:06 am | Permalink

      I’m confused.Wouldn’t that be the opposite,that the nDNA goes back farther than 200,000 yrs?

      • Posted July 27, 2012 at 6:01 pm | Permalink

        I would like to know if this study has any relevance to the “Mitochondrial Eve” story in humans. As I remember it (just as a layperson watching TV shows…) there is some disagreement in the paleoanthropological community about divergence times for the various varieties of modern humans, with mitochondrial evidence showing a more recent divergence seeming to conflict with anatomical fossil evidence seeming to show a more distant divergence.

  8. Newman
    Posted July 24, 2012 at 7:03 am | Permalink

    Jerry, if you publish this for realz, I will cite you in my dissertation proposal. heh.

    This is an interesting paper. Nothing new (as far as methodology) for those of us in the field of systematics and phylogenetics, since there are many recent papers showing such cytonuclear discordance as presented here. Still, the fact that these animals are charismatic species make this particular study newsworthy, and hopefully it will provide a needed push toward use of genomic data for those who may still be hesitant.

  9. phosphoros99
    Posted July 24, 2012 at 8:05 am | Permalink

    Thanks for a very informative post.

  10. Occam
    Posted July 24, 2012 at 8:37 am | Permalink

    Marine Isotope Stage 11, the interglacial with the sharpest decline of Polar bear population in Fig. 6, is generally viewed as the closest interglacial analogue to the Holocene. And I just read (Geophysical Research Abstracts Vol. 14, EGU2012-2362, 2012) that the CH4 (methane) increase during MIS 11 has been shown as very similar to the present.
    Bad news for PB. And for us. We don’t deserve this planet.

  11. Posted July 24, 2012 at 8:41 am | Permalink

    Wait a second. We now learn that the polar bear evolved 3 million years ago. That was near the beginning or before the onset of the current ice house. Therefore, it survived not only the FIRST glaciation, but 17 cycles of glaciation/interglacial. The bear has known nothing but stress, change and the need for migration.

    At one end of extremes, there was little or no arctic sea ice (as for a thousand years during the Holocene Climate Optimum) and at the other huge ice sheets a mile thick down to Chicago. Seventeen times.

    Through it all, the polar bear survived. Even if AGW causes a delay of the next glaciation, even if the arctic ice fades to nothing for 1000 years until the glaciation can no longer be denied, what is the basis for your claim that “…polar bears are sensitive to climate…those bears are liable to extinction…unable to support themselves ecologically as the polar ice disappears…”

    And I might as well throw the other boot across the room: “…lovely animal…” is in the eye of the beholder. This is a savage enraged beast that must keep cubs away from males for fear they will eat them, and nourishes by discovering vulnerable seals and eating the seal babies alive in front of the seal mothers. I think they are unlovely.

    • whyevolutionistrue
      Posted July 24, 2012 at 8:52 am | Permalink

      I simply speculated that, since the bears’ diet is seals, their inability to get to the seals permanently, as might occur with global warming, might lead to their extinction before they can adapt to a new environment or diet.

      Your disdain for the beauty of the bears is duly noted. I presume, then, that you think tigers, lions, cheetahs, and other carnivores are “unlovely.” Well, that’s your opinion. Natural selection does what it does.

      • gbjames
        Posted July 24, 2012 at 8:57 am | Permalink

        Maybe, like Stephen Colbert, John thinks of bears simply as godless killing machines.

      • Posted July 24, 2012 at 9:26 am | Permalink

        So how did the bears get to the seals during a) 17 interglacials when there was little or no sea ice; and b) during 17 glaciations when the ice sheets were extended?

        Also, what do you mean by “permanently?” The end of all ice forever in the North?

        • Posted July 24, 2012 at 4:50 pm | Permalink

          My understanding is that there was some sea ice sheets during interglacials; for example, we are in an interglacial right now and there is sea ice and polar bears. And of course, polar bears didn’t evolve instantaneously 3 million years ago. Their ancient population probably adapted to the shifting conditions over time, and we don’t know if a population, say, 2.3 million years ago or 400,000 years ago, had similar adaptations (e.g., foraging strategies, diet, coat color) as current polar bears. The key issue today is whether polar bears (or tigers or humans for that matter) can adapt to changing conditions based on human-induced climate change that happens on a much faster scale than the recurrent glacial-interglacial periods. Creatures with long generation times are likely to be in trouble.

          As for polar bears being sensitive to climate change, these types of statements are based on demographic projections. Projections are different than predictions. Projections state things like, if X happens, then Y will follow. In terms of polar bears, several top demographers got together to do a study for the US Geological Survey; their conclusion–which included modeling population growth rate as a function of future sea ice availability (using IPCC models)–was that if current population growth rates remain the same and sea ice declines, the polar bear will decline to extinction.

          • Posted July 24, 2012 at 10:16 pm | Permalink

            “My understanding is that there was some sea ice sheets during interglacials;”

            Check your facts. It is quite possible that the only ice in the north for long stretches of interglacials, including for 1000 years in the Holocene, was the Greenland sheet. No or very little sea ice.

            The bears adapted to this many times.

        • gravelinspector
          Posted July 25, 2012 at 4:28 am | Permalink

          Also, what do you mean by “permanently?” The end of all ice forever in the North?

          A significant difference from previous interglacials is that in the coming global warming the polar bears be migrating in the presence of a superior climax predator. Which uses rifles. Rifles will have a significant effect on the success of their attempted migrations.
          OTOH, I bet that someone, somewhere, for some motive, plans to release polar bears into Antarctica. Which is going to be “lovely” for the penguins. Maybe it would be sensible to see how penguins survive in the Arctic first, but that’s likely to take too long for some people.

      • Posted July 24, 2012 at 9:33 am | Permalink

        Natural selection does what it does. Well, then your “lovely” is an undeserved emotional privileging by wont of dis-inclusion that you think all life is lovely. The killing is lovely. The nurturing of young is lovely. The repeated stupidity of herds crossing in the same place crocs gather every migration is lovely. Elephants mourning is lovely. Chimps slaughtering rivals and eating them is lovely. Penguin males nesting all winter with an egg between two feet is lovely.

        The word has lost all meaning.

        • darrelle
          Posted July 24, 2012 at 11:14 am | Permalink

          An absurd response. The OPs meaning and context was clear. What is it that is really bugging you?

          • Posted July 24, 2012 at 10:19 pm | Permalink

            A scientist proffering emotional propaganda at the bottom of a factual essay, that’s what’s bugging me.

            • gbjames
              Posted July 25, 2012 at 4:20 am | Permalink

              Propaganda? Srsly? In service of those godless killing machines?

              Whoo boy. You need to install a new sense-of-perspective module.

        • Occam
          Posted July 24, 2012 at 11:44 am | Permalink

          “Emotional privileging” is by definition undeserved.

          Our emotional response to various species may be in part mediated by culture, but I presume it is mostly a by-product of evolution.
          There may be many reasons, some of them evolutionary, why many humans show an emotional response to certain species, unless they happen to live in direct contact with them. And maybe more so if they do.

          I spent long periods of my childhood on sheep farms in the Carpathian Mountains. Seeing sheep killed by wolves, or by the occasional bear, was not exceptional. I’m glad I had this early a chance to observe wolves and bears in the wild, from afar, for I learned to respect and admire them. I knew they could kill me under the wrong circumstances; I learned how to behave to minimize the risk. They’re splendid beasts, I wouldn’t mind calling them lovely, though I’d rather reserve the adjective for their cubs. Even if I recall quite vividly what a sheep cadaver with its belly torn open by wolves looks like. They’re predators, not plush toys. The acceptance of biological reality does not preclude a positive emotional response.

          On the other hand, many people feel that cats are lovely. I don’t (Caturday felids are an acquired taste, thanks to JAC). When I was a kid, I was attacked by a lynx. The physical damage was slight, the feline trauma persists to this day. Again, an “undeserved” emotional response, this time a negative one — blame natural selection for that.

          BTW, your Humpty-Dumpty-ing with lovely shows quite conclusively that the word still has a meaning: either the customary one, or the one you assign to it.

          • Posted July 24, 2012 at 10:10 pm | Permalink

            @Occam
            “…culture…evolution…” etc

            I reject your rejection of personal responsibility for value choices such as “lovely.”

            This person chose his privileging. If he did not consciously choose it, then he let some default propaganda do it. Still it is his choice and personal responsibility. He chose the saccharine cutesy image of the beast instead of the vicious ones where the white fur is soaked with blood and seal intestines litter the ice. In this particular case, power advertising by Coca Cola, with stupendous pile-on by the perfidious Al Gore and the propaganda of Adam Ravetch et al., is in place to shove a value in if you choose not to.

            My point about “lovely,” which none of you are bright enough to grasp, is that if you apply a human positive value such as “lovely” on a predator that does nothing but kill helpless babies, you devalue the concept behind the word.

            To clarify my own position, the actions, fate and interplay in all of metaphysically given reality, including living flora/fauna, has a stark beauty for the human soul simply because it is real, the only real there is. Honoring it with respect and clear unclouded vision is the beautiful act of a lovely species: homo sapiens.

            • darrelle
              Posted July 25, 2012 at 6:52 am | Permalink

              “My point about “lovely,” which none of you are bright enough to grasp, is that if you apply a human positive value such as “lovely” on a predator that does nothing but kill helpless babies, you devalue the concept behind the word.”

              I wouldn’t be so quick to hurl gratuitous insults about other peoples intelligence while simultaneously making such ridiculous comments. Your comment amounts to nothing more than an opinion about something which is largely subjective, you engage in hyperbole, you assume facts not in evidence and ignore much evidence against your claim. You also display a poor understanding of context, including social context.

              So, what has you so emotionally engaged that in your haste to smack people down you make such silly comments? Religious beliefs? Climate change denialism of some stripe? Political ideology?

              • Posted July 25, 2012 at 5:26 pm | Permalink

                ok you reacted to my “bright enough” phrase and threw four times the insults back without justifying them. I withdraw that phrase and the rest stands.

                My point about “lovely,” is that if you apply a human positive value such as “lovely” on a predator that does nothing but kill helpless babies, you devalue the concept behind the word.

                Do you care to actaully respond to my post now?

            • Posted July 26, 2012 at 7:40 am | Permalink

              Actually, I would like to extend a general apology for the “bright enough” phrase and withdraw it. Yes, I am angry about the exploitation of the PB for political reasons, but that comment was rude and I am sorry I posted it.

        • CJ
          Posted July 24, 2012 at 1:46 pm | Permalink

          You might as well give up on life Johnny boy. Because you were born of destruction. You wouldn’t be here if stars hadn’t exploded, and parasites and viruses hadn’t flourished. Nor would all that great music. You cannot have the Yin without the Yang. And that is lovely!

          Just because an animal is dangerous and revolting in one context, doesn’t mean it is in all others.

          “You are what you love, not what loves you”
          Donald (a.k.a Charlie) Kaufman.

    • Torbjörn Larsson, OM
      Posted July 24, 2012 at 1:54 pm | Permalink

      It is interesting that the polar bear evolved ~ 4-5 million years ago, when, arguably, both polar caps evolved ~ 15 million years ago, well before the glaciations, due to the cooling at the end of the Neogene.

      That indicates that the lifestyle is hard to evolve.

      The paper notes that the polar bear has been in decline for a long period over many of these climate changes. The current one, removing more ice than heretofore and possibly _all_ ice for the first and last time, could well be the undoing of this species.

      At the same time the beauty of this survivor of the extreme, bred to relative gigantism in a food scarce and cold environment, is to manage to be the necessary top predator for a balanced and diversified ecology. If it goes, so goes the ecological environment for many species.

      I see your natural fallacy and raise it: parasites are the most numerous of all life styles. Apparently nature loves parasites.

      • Posted July 24, 2012 at 10:29 pm | Permalink

        I cannot find a paper that substantiates that Northern Sea ice of any substantial extent existed earlier than 2.4 million years ago. Can you cite? I see studies of cores off Greenland [Thiede, JC Jessen et al.] that substantiate some ice on that land, perhaps glaciers, earlier.

        If not, and the date for the startup of north sea ice is indeed in the 3MYA window, that is a correlation for the simultaneous evolving of the polar bear. Interesting. That in itself does not prove causality. However, combined with the obvious color adaptation I would think amounts to a smoking gun.

        Second, with regard to “possibly _all_ ice for the first and last time”, obviously you are asking for a contextual stipulation on “first” to refer only to the current brief 2.4MY ice age, since for most of earth’s history there was no ice. Granted. As to “last”, not many scientists would stake their reputation on a belief that burning some of the current carbon deposit trumps, absolutely trumps, the massive forces that drive the current Ice Age.

    • raven
      Posted July 24, 2012 at 2:07 pm | Permalink

      and nourishes by discovering vulnerable seals and eating the seal babies alive in front of the seal mothers. I think they are unlovely.

      Humans kill and eat adult and baby cows, pigs, chickens, fish, lobsters, and, in fact, just about any animals that aren’t poisonous by someone, somewhere. The habit of eating embryonic chickens aka eggs, is particularly weird

      I assume you think such indiscriminate slaughter by those omnivorous humans with a carnivorous bias are equally “unlovely”.

      But don’t worry. They think you are “unlovely” right back. What a kook.

      • gravelinspector
        Posted July 25, 2012 at 4:32 am | Permalink

        “Seal babies”, “cow babies”, “embryonic chickens”. Yummy! Bring on the “Modest Proposal”, I say!
        Fricassee, anyone?

  12. Maria Jinich
    Posted July 24, 2012 at 9:07 am | Permalink

    What about ancestry analysis with mtDNA? Is it reliable? Like the one done by NatGeo or 23andme.

    • The Other Jim
      Posted July 25, 2012 at 1:16 am | Permalink

      It reliably tells you about your maternal heritage. As it did in the bear study. But there are some inferences that need to be made with caution, as this illustrates.

  13. Jim Thomerson
    Posted July 24, 2012 at 9:13 am | Permalink

    I have wondered about why mitochondrial genes introgress more readily than nuclear genes. The idea that hybridization events usually involve females of the rarer species seems reasonable. This would suggest that mitochondria introgress more readily into the genome of the more abundant species.

    • Gregory Kusnick
      Posted July 24, 2012 at 12:40 pm | Permalink

      I’m not an expert, but my guess is that it’s not so much that they introgress more readily, but that they don’t undergo recombination or dilution. This creates an observational bias in which we see only the ones that successfully go to fixation. The ones that don’t get eliminated wholesale and leave no trace that they ever existed.

      • Torbjörn Larsson, OM
        Posted July 24, 2012 at 2:01 pm | Permalink

        Why would they not undergo dilution? I thought mtDNA was a lot of assorted mitochondrial genomes per cell. Drift would be paramount, I take it.

        • Gregory Kusnick
          Posted July 24, 2012 at 2:42 pm | Permalink

          The point is that any given bear carries either PB or BB mtDNA, never a mix of the two. So the matrilineal dynasties of PB and BB mtDNA must remain forever separate until one of them dies out. Unlike nDNA, there’s no way for mtDNA to escape extinction by mingling with the other kind.

    • van Rooinek
      Posted July 25, 2012 at 11:30 am | Permalink

      I have wondered about why mitochondrial genes introgress more readily than nuclear genes.

      One possible reason is Haldane’s rule. When 2 species start to diverge to the point where interbreeding leads to sterility, the heterogametic sex (males in mammals) is the first to exhibit hybrid sterility.

      For example: cattle x North American bison, the male hybrids, “cattalo”, are sterile, but the female cattlo are fertile and can be backcrossed to either parent species. Many N.A. bison herds today, show some cattle mtdna here and there, indicating that female cattle introgressed.

      The horse/donkey genetic separation is greater –not only the males but also nearly all female mules are sterile. Yet, on rare occasions a fertile female mule occurs and can backcross with a male horse or donkey. So in theory, if we looked, we might find phenotyic donkeys with horse mtdna from a long forgotten fertile-mule mother.

      The bottom line is, that mtdna can jump past a partial sterility barrier. But then that female has to backcross with a pure member of one of the parent species. With each backcross the introgressive nuclear DNA is diluted ~50%, and if crosses are rare, it will eventually be undetectable. But the introgressive mtdna is passed on intact.

  14. Posted July 24, 2012 at 10:23 am | Permalink

    It would help if this post explained how mtDNA leaks from one species to another. Here’s my guess, given that it was mentioned above that mtDNA comes from mothers only:

    A female PB and male BB mate and give birth to a fertile female hybrid. The hybrid has a mix of PB and BB nDNA, but the mtDNA is from the PB alone (since she was the mother). The hybrid then mates with another BB male, resulting in a hybrid with a majority of BB nDNA, but full PB mtDNA.

    Is this accurate?

    • chascpeterson
      Posted July 25, 2012 at 8:47 am | Permalink

      it’s backwards, I think. Brown bear mDNA introgressed relatively recently into the polar bear gene pool.

  15. Posted July 24, 2012 at 11:02 am | Permalink

    Jerry, is the hybridization mentioned the same thing as coalescence?

    I’ll have to go back and look at PZ’s posts about it, but it sounds similar.

    Thanks for a great post!

    • Josh
      Posted July 24, 2012 at 2:24 pm | Permalink

      Lynn,

      Coalescence is actually a more general thing: it just refers to when two individual from a species share a common ancestor at a given place in their genome. If we look at say, nucleotides 135,303,302-135,303,500 on chromosome 1 of your genome and my genome, at some point in the past we would have shared a great-great-great (etc.) grand parent and both of us would have gotten that segment of our chromosome from that person. In that case, we would say that that portion of our genome coalesced in that ancestor.

  16. Josh
    Posted July 24, 2012 at 12:04 pm | Permalink

    Jerry, you need to write a strongly worded paper to get people to stop using mtDNA. I feel like you could have the right kind of impact!

  17. Posted July 24, 2012 at 12:40 pm | Permalink

    This is interesting as far as it goes, but I think it would be even more interesting and useful if the study included all the living Ursids. For example, how closely related is the American black bear to the Tibetan moon bear or the Asian sloth bear? Also, sorting out these relationships might offer genetic clues about the extinct Pleistocene cave bear and short-faced bear.

    Zoos, conservation organizations, paleontologists, and even ordinary citizens who are into bears might help support a basic, essential research project like this.

    • Posted July 24, 2012 at 1:14 pm | Permalink

      I think that would be a bear necessity.

      /@

      • Torbjörn Larsson, OM
        Posted July 24, 2012 at 2:04 pm | Permalink

        As long as your constraints don’t polar-ice the issue.

  18. Alex SL
    Posted July 24, 2012 at 1:01 pm | Permalink

    Well, that about the chloroplast and mitochondrial capture should (one hopes) be widely known in the community – but often sequencing those regions is still the easiest thing to do. I mean, not every honours student can do cloning and Sanger sequencing on ten nuclear genes or Next Gen Sequencing of whole transcriptomes, and especially not for 80 samples or so. You may more easily get that money for polar bears or gorillas than for some plants of no particular commercial interest. Luckily, in botany the chloroplast is also fairly slow-evolving, meaning that it is most useful at higher taxonomic levels where introgression should be rarer.

    While you have touched on it in the end, I think more attention should be given to incomplete lineage sorting. We can have discordant gene trees even without any introgression, we even expect them in the first time after speciation unless it happened through a severe bottleneck like one seed being blown onto an island. The solution is, of course, again the expensive one to examine many nuclear loci.

    If that were true it would mean that the species “brown bear” was “paraphyletic,” i.e., some populations of brown bears are more closely related to another species than to other brown bears. For some biologists, that menas that brown bears wouldn’t be qualified as a “species”.

    I know of those colleagues, but really “this species is paraphyletic” makes about as much sense as “the national anthem tastes of strawberries”. Phylogenetic systematics deals with identifying monophyletic groups of species. Within sexually reproducing species, the structure is predominantly tokogenetic, not phylogenetic. That is kind of the point.

  19. Torbjörn Larsson, OM
    Posted July 24, 2012 at 1:14 pm | Permalink

    The similar Neanderthal vs Modern mitochondria history may pop up in the mind of the layman. Based on mitochondria thus far, Neanderthals never crossed back with Moderns.

    The problem is that, for reasons we don’t fully understand, mtDNA also moves between species during hybridization much more readily than does nDNA, and that can screw up species relationships.

    To me it looks as a case of “use at least hundreds of genes, preferably random” if not near complete genomes are feasible. (They don’t have to be the perfect, well resolved complete genome either.)

    But that mtDNA and cpDNA is readily movable I naively understood as a result of successful endosymbiosis in the first place. By moving as many necessary genes as possible to the nDNA respectively trashing the unnecessary genes would improve success rate.

    Not my business, but I would tag that DNA as “peculiar”.

    How it happens is therefore intriguing though. Curious laymen want to know! Do we have so few enterprising biology graduate students? (And if so, why and how? Pesky science, always exploding the question set!)

    • Posted July 24, 2012 at 1:22 pm | Permalink

      I think it would be cool to become a biology student when I retire.* 🙂

      /@

      * ca. 13 years away

  20. gillt
    Posted July 24, 2012 at 2:03 pm | Permalink

    another reasong why mtDNA is bad: paternal leakage

  21. Barbara
    Posted July 24, 2012 at 5:03 pm | Permalink

    When all we had for DNA phylogeny of polar bears was the mtDNA and some grizzly/brown bear populations were found to be more closely related to polar bears than to other grizzly/brown bears, the grizzly/brown bear could be considered paraphyletic. Some people think paraphyly is horrible and forbidden, and therefore argued for inclusion of polar bear within the grizzly/brown bear species. Conveniently, the grizzly/brown/polar bear would not be rare enough to receive stringent legal protection.

    That frustrates me because one would EXPECT a parental species to be paraphyletic to a newly evolved species. In widespread species A, certain populations at the edge of its geological/ecological range become isolated. Due to drift and/or selection (sometimes with polyploidy thrown in), those populations become reproductively isolated, forming species B. However, they’re still more closely related to the populations located near species B than to those a continent away. Species A is paraphyletic to B, but B is still be a good, reproductively isolated species. Eventually, species A and B would likely change so much that the similarity of B to the nearest A populations would be obscured, but in the mean time we might notice the paraphyly. And it doesn’t matter.

    I’m glad that there is now good nuclear DNA evidence for the great depth of the split between polar and grizzly/brown bears. However, the whole argument based on paraphyly was misguided all along. Polar bears differ from grizzly/brown bears morphologically, ecologically, and (mostly) geographically, and they rarely interbreed. That’s about as good a species-level difference as we are likely to see.

  22. pjlandis
    Posted July 24, 2012 at 9:44 pm | Permalink

    Is Mitochondrial DNA separate from the main DNA in a blastocyst? Or is mDNA written into nDNA?

    • The Other Jim
      Posted July 25, 2012 at 1:19 am | Permalink

      It is physically distinct from the nDNA, existing in the mitochondria, within the cytoplasm of the cells. The oocyte that made you carried 1 copy of each nDNA region, but ~200,000 copies of the mtDNA.

  23. Posted July 25, 2012 at 1:25 am | Permalink

    Yep 🙂

    Long-winded version:

    Mitochondrial DNA is found in mitochondria. You inherit your mitochondria from your mother’s gametes alone, and you inherit them as whole, intact organelles. In animals, “eggs” contain all the organelles needed by a new zygote; “sperm” contains basically nothing except nuclear DNA (so no mitochondria at all).

    Nuclear DNA is found in nuclei. Animals (typically*) get half of their nuclear DNA from their father(‘s sperm), and half from their mother(‘s eggs).

    The two types of DNA don’t directly physically interact with each other. The nucleus and the mitochondria are separate organelles within a cell, separated by membranes across which the DNA doesn’t move (except during cell division in the case of the nucleus).

    During cell division, the nuclear membrane is dissolved, nuclear chromosomes (i.e. connected strands of nuclear DNA) duplicate themselves, and then split apart: one of each identical pair goes into each new cell. A new nuclear membrane forms around the set of “nuclear” chromosomes in each new cell.

    Mitochondrial membranes, on the other hand, don’t dissolve during cell division: after duplicating their DNA, the mitochondrial membrane pinches in the middle to split the mitochondria into two new duplicates, one of which ends up in each daughter cell. Mitochondrial DNA doesn’t (ever?) leave the confines of its mitochondrial membrane.

    During meiosis, you also have a “crossing-over” step between (typically) non-identical (i.e. paternal & maternal) pairs of nuclear chromosomes to generate new combinations prior to further chromosome replication and cell division. Mitochondrial DNA doesn’t do this (no non-identical ‘pairs’ to participate in an exchange). This is what makes mDNA so handy in tracing maternal lineages: there’s no ‘noise’ produced by crossing-over, no blending of genomes, the only changes over time are mutations. Unfortunately, it also causes problems for species-level analysis, as you lose a lot of information when a single maternal mitochondrial line spreads (whether by drift or because of a selective pressure) through a population.

    *In fact, feel free to insert “typically” anywhere in this comment. There are always exceptions…

    • Posted July 25, 2012 at 1:26 am | Permalink

      Er… that’s supposed to be a reply to pjlandis @#22. Sorry about the blip :/

    • chascpeterson
      Posted July 25, 2012 at 8:52 am | Permalink

      “sperm” contains basically nothing except nuclear DNA (so no mitochondria at all)

      actually sperm cells do have a mitochondrion, but it (almost) never enters the ovum.

  24. chascpeterson
    Posted July 25, 2012 at 8:57 am | Permalink

    Can somebody who’s read the two 2012 papers address an apparent discrepancy? AFAICT from second-hand sources like this, one puts the polar/brown bear split at 600,000 ya and the other one at 4-5 million ya. That’s a big difference.

  25. mmp
    Posted July 25, 2012 at 9:37 am | Permalink

    That’s probably a newbee question, but could it be that mtDNA does not give the correct phylogeny (for bears) because the dispersal sex is female (and mtDNA is maternally inherited)? On that note, couldn’t it be that, in animals whose dispersal sex is male, we have the opposite situation: nuclear genes give us the incorrect phylogeny while mtDNA give us the correct one?

  26. Posted August 5, 2012 at 6:20 am | Permalink

    Another example of when mtDNA based systematics can go awry is when migration behavior makes females of a species gather for mating and return to their origin while the males follow the females regardless of their origin. In these situations, like are seen in some Australian and Holarctic duck species, you will get two distinct and isolated mitochondrial genomes, but the nuclear genomes will be indistinguishable. This calls into question not only the singular use of mtDNA in systematics, but also the “barcode gap” used by some in identifying cryptic species. Male dispersal and female tendency to stay put is seen in many groups of animals, so I wouldn’t be surprised if this issue is widespread.

    So, to answer one of the above questions, mtDNA might more likely give us the “correct phylogeny” (as if phylogenies should ever be based on a single character rather than the congruence of multiple characters), but they will also wrongly multiply the number of isolated lineages we see.

  27. Luke Reding
    Posted October 14, 2012 at 7:43 am | Permalink

    Mitochondrial DNA is preferred when using a coalescent approach because a) mtDNA is inherited via the matriline and b) mtDNA doesn’t recombine like nuclear DNA. Recombination generally leads to significant errors when estimating a coalescent event, so I think there are legitimate reasons for using mtDNA in addition to nuclear DNA. In the 2010 study you site, the authors admittedly don’t take a coalescent approach.


3 Trackbacks/Pingbacks

  1. […] A new study of polar bears underlines the dangers of reconstructing evolution using mitochondrial DN… (whyevolutionistrue.wordpress.com) […]

  2. […] Polar bears are older than we think, and Jerry Coyne has an excellent discussion about the dangers of using only mitochondrial DNA for constr…. […]

  3. […] This discovery highlighted the perils of constructing phylogeny and taxonomy using only mtDNA data. […]

%d bloggers like this: