Posts Tagged phylogenetics

Sunday Spinelessness – 5 down…. quite a few to go David Winter Jan 20


I got some good news this week – a paper I’m an author on was accepted for publication pending some minor revisions. That’s great because career advacement in academia rests largely on what we publish, and this is a good paper that I’ll be happy to add to my CV. It’s also quite happy about his particular paper being (almost) accepted because it’s about serpulids, segmented worms of the phylum Annelida (relatives of earthworms). A new phylum for me.

Biology is about diversity. I know I always go on about this, and end up affecting the overly-enthusiastic style of the guide in Douglas Adams’s Hitchiker’s Guide to the Universe:

Biological is diverse. You just won’t believe how vastly, hugely, mind- bogglingly diverse it is. I mean, you might think there are lot of creatures in your average David Attenborough documentary, but that’s just peanuts to the true diversity of biological systems, listen…

Well, I don’t know to put in words, so let’s try a picture. All that biological diversity got here because life evolves. When populations break up they are free to evolve apart from each other and develop entirely new functions or features and so become different. In this way, life is a tree, forming new branches as populations split. When we come to deal with the diversity of life, biologists try to reconstruct that tree, giving names to those tips and twigs which belong to a particular branch. In that  system of classification the phylum (plura phyla) is the one of the deepest divisions.

 Creatures in separate phyla have usually been evolving apart from each other for 600 million years or more, and represent entirely different ways to deal with the trials of life. The annelid paper will mean I’ve published on 5 different phyla. That’s exciting for me – it’s nice to think I’ve added a little to our knowledge of decent sampling of the tree of life. But the truth is, biology is just so diverse that I’ve not even made a dent the tree of life. Here’s a picture of all the Eukaryotic phyla (that is, creatures with cells like ours, but not bacteria and archaea) with only those I’ve published at least one paper on labeled:

Tree was drawn and shaded with iTOL‘s nifty interfact to the NCBI taxonomy. There’s a couple of things to note here. Because this is the NCBI taxonomy it’s a curated tree rather than the result of any particular analysis. Although we aim to create biological groups “natural”, in the sense they are a single branch in the tree of life, the rank giving to a particular branch is somewhat arbitrary and will differ between different groups (so green plants, which traditionally had “divisions”  rather than phyla are certainly underrepresented here). Protists (single-celled eukaryotes) are certainly diverse, but Psi Wavefunction tells me protistologists have almost given up on rank-based taxonomy so this might not be a fair representation of them.
In any case, it’s certainly a spur to me to get back to work and fill in a few blanks on the figure!

Chimps are our closest relatives… but not for all of our genes David Winter Mar 15

Ladies and genetlemen, we have the gorilla genome. You reaction to this news is probably determined by what you do for a living. If you write the headlines for major news services you will convince yourself that this result will, in some utterly undefined way, teach us about what it is to be human. Just about everyone else will develop a case of  Yet-Another-Genome Syndrome. The gorilla is, by my count, the 51st animal to be added to the full-genome club and the last of the great apes (joining humans, chimps and orangutans). More to the point, the publication of a new genome sequence doesn’t, by itself, tell us all that much. The real achievement in a “genome of x” paper is the creation of a resource that scientists will continue to work from for decades. The analysis that comes with it is really just a first pass.

But there was one very cool result to come from the analysis of the gorilla genome. About 15% of our genes are more closely related to their counterparts in gorillas than they are to the same genes in chimps.

That sounds suprising. People are always going on about how humans and chimps are ninety-nine-point-some-magic-number percent identical, and there are exactly two scientists in the world who think chimps are not our closest relatives (Grehan and Schwartz, 2009 doi: 10.1111/j.1365-2699.2009.02141.x). Have we been wrong? And how can 15% of a genome show one pattern while the rest shows another?

To understand what’s going on, we need to remember where species come from. Species start forming when populations stop sharing genes which other. When genetic changes in one population can’t filter through to another, those two populations are capable of evolving apart from each other and so can become distinct and take on the various characters that we use to tell species apart. So, new species only become different as they start to evolve apart, but they start of with a more or less random sampling of the genes in the ancestral population from which they descend. If we want to understand what’s going with the gorilla genome, we need to understand the history of those genes.

In most populations at least some genes come in distinct “flavors” (technically called alleles) . So, for instance  we all have a gene called MC1R, but some of have an MC1R allele that is associated with red hair, and others have alleles that usually lead to dark hair. We inherit  our genes from our parents, so each allele has a history that stretches back through time. If we look at modern populations we can use genetic differences between alleles to reconstruct that evolutionary history. Here’s a simplified history of four alleles, in a very small population (if you re-trace the lineages you see they fit the tree to the right):

So, what happens when a population with different alleles starts to diverge into new species:

The genetic lineages will keep on evolving down through the new tree, but now lineages will never cross the barriers to gene flow that are driving speciation. Often, the genetic lineages in the ancestral population will “sort” in such a way that when you trace the genetic lineages within a species back you arive at a member of that species (not an individual from the ancestral population). In that case, the genetic relationships (which we’ll call “gene trees) will be the same as relationships between species (“species trees”):
But population genetic theory tells us we won’t always get such a simple pattern. For recent or repeated and rapid speciation processes there might not be time for the genetic lineages to sort. The gene tree can be different from the species tree:

Exactly this process has happened with the gorilla genome. The genetic lineages hadn’t sorted before the human-chimp split so some of our genes are more closely related to gorilla ones than chimps ones.  This phenomenon might tell us something about the evolution of the great apes . The time that it takes for lineages to sort is proportional to the population size of the organisms through which the lineages are evolving. Processes that effectively limit the population size (like natural selection, which results from relatively few individuals contributing to the next generation) might leave a pattern in the way lineages have sorted.  The authors of the gorilla genome paper use this prediction to search for and find areas of the gorilla genome that may have been subject to strong selection after the population went its own way.

So called “incomplete lineage sorting” is a problem for people like me who aim to reconstruct the evolutionary history of species using genetic data. Although we’ve always known this problem existed, we’ve only recently been able to extend population genetics theory to actually infer the history of species for gene trees even when those gene trees are unsorted. It’s important we have these methods, because it’s actually predicted that most genetic lineages will be unsorted for about 1 million years after speciation starts – often all we have are unsorted genes and it’s nice to be able to extract some information from them.

The Gorilla Genome paper is

 Scally, A., Dutheil, J. Y., Hillier, L. W. et al. (2012) Insights into hominid evolution from the gorilla genome sequence. Nature, 483, 169-175 doi:10.1038/nature10842

The Tree of Diversification (or why the March of Progress is wrong) David Winter Feb 22


I’m giving a big talk next week – a departmental seminar. It’s the first time I’ve had more than 15 minutes to talk about my research, so this talk will be a little more discursive that my usual.

I study speciation, how new species come into being, and one of the things that I want to emphasise is that speciation hasn’t really entered into the broader understanding of what evolution is.Take the one image that describes evolution in modern society:

The March of Progress, Rudolph F. Zallinger. From Time Life’s book Early Man.

The so called “march of progress” has been used to describe the origin of our species thousands upon thousands of times. But it never happened. Only a few of the species depicted are potential ancestors for humans and many of them were contemporaries to each other (as the original diagram makes clear) so can hardly be different steps along a single evolutionary path. 
To try show what really happened, I’ve redrawn “March of Progress” into the “tree of diversification” – trying to show how the species depicted above relate to each other (parts of this tree are very much up for debate, by the way). Bear in mind, I’m only including the species that are represented in the original graphic, if we were to include all the fossil ape species we know about the tree would be much bushier):

Silhouettes are CC BY-SA by José-Manuel Benitos this image is released under the same license

I think that when you look at it this way it becomes clear that if we want to understand how the organisms depicted in the most famous icon of evoluton came to be we need to focus not just on how changes occur in one lineage, but on how new lineage form and become capable of changing in their own directions. At the moment there are probably 10 million species on earth, and just how they got to be here is surely one of the biggest questions that biology asks us. Speciation and diversification ought to be central to the way we think about evolution.

Sunday Spinelessness – The first animals (molecular biology) David Winter Jul 10


It’s time to wrap up this series of posts on the origin of animals. If you are just tuning in here, I’ve already decided that early fossils are utterly fascinating but really not much help to us and had a look at some modern organisms that might give us an idea about how the major steps towards multi-cellularity might have been achieved. Today, I’m going to zoom down another level, and see if molecular biology can tell us anything about the first animals.

Life is really just a very special kind of chemistry. The organisms that emerge from the bewildering number of inter-connected chemical processes that make life tick are so complex, so varied and so far removed from the things that chemists study that we had to create a whole other science, biology, to study them. For a long time biologists and chemists more or less got on with their own problems, but in the middle of the 20th century the progress each field had made created the opportunity to actually study the molecules that make life. The so called “molecular revolution” in biology has effected almost every sub-discipline within our science, and the tools and data that revolution created can be used to try and understand the origin of animals. Over the last decade or so scientists have sequenced the genomes of an increasingly number of species, and in between the big headlines created by creatures like chimps, mice and humans; we’ve been able to learn about sponges, cnidarians and placozoans. Those sequences can help us understand how these animals work at a molecular level, and they can also be used to recover the relationships between different groups of animals.

Scientists call the branching trees we create to represent the relationships between species ‘phylogenies’, and as someone who spends quite a lot of his time working on molecular phylogenies (have you seen The Atavism’s logo…) I should really be saying that this is the best way to understand the origin of animals. But that’s not really true. DNA isn’t quite the wonder-molecule that CSI may have lead you to believe it is, and having DNA sequences available isn’t necessarily enough for us to recreate a phylogeny. At the moment, the relationships at the base of the animal family tree are very hard to untangle. We know that a group of animals called the bilatarians (you, me, insects, fish, sea urchins, nematodes, spiders and, well, really every animal you are likely to name off the top of your head) form one group distinct from the other animals, and that these animals can be further subdivided based on the way their embryology plays out. But the relationships of the non-bilatarian animals – a grab-bag of creatures including the cindarians (jellyfish, corals and their relatives), comb jellies, placozoans and sponges – are really unclear:

That uncertainty isn’t a matter of us just not having enough data to throw at the problem. Molecular phylogeny works by grouping species together based on mutations that they share, having inherited them from a common ancestor. So, ideally all the branches in the true history of the species we are studying will be separated by enough time for those mutations to start racking up in the period of time that lineages shared an evolutionary history, something like this (with the mutations being the red stars occurring at random):

Sadly, the real pattern underling the base of the animal family tree doesn’t appear to be like this. The initial branches fromed relatively quickly, leaving only short periods of shared ancestry for related groups to accrue mutations that would allow us to recreate their relationship. Worse yet, the branches that lead out from initial splits are really long, which means there has been millions of years for any useful mutations to have been over-written by more recent changes in DNA sequences.

There is no easy way to get around this problem. It seems like we should be able to throw more and more genetic sequences at the tree, increasing the chance the we get some informative mutations, but these trees run the risk of falling into the terrifying Felsenstein zone, in which the confidence with which we estimate the wrong tree increases as we apply more sequences to the problem. Because of the difficulties inherent in reconstructing the very early history of animals, I think we should take results based solely on molecular phylogenies with more than a few grains of salt.

A case in point is the phylogeny Casey Dunn and colleagues presented in Nature a few years ago (doi: 10.1038/nature06614). If we were trying to fill in the question marks and the base of the animal tree without considering molecular data at all, we’d probably put sponges on the first branch to split in the animal family tree. Sponges are quite different from other animals, they don’t have a nervous system, their cells are not bound into tissues and, although sponges do have specialised cell-types, their cells are uniquely flexible in that mature cells can change from one type to another. It’s actually quite possible that several of the early branches in the animal tree lead to different types of sponges, like this:

But, against all expectations, Dunn’s phylogeny found comb jellies (creatures that look a little like jelly fish but form a phylum, Ctenophora, in their own right) to be the first group to split from the animal tree. If that were true, then we’d be less confident that the first animals were sponge-like since either (a) sponges would be radically modified animals which had lost their nervous system, given up on tissues and taken to a sedentary filter feeding life or (b) comb jellies would have evolved nervous systems and tissues independent of those of other animals.

There is no reason that such a scenario would be impossible (in fact, fungi and plants have each evolved tissues independently of animals), but when the whole idea is constructed from a method that we know is prone to biases I don’t think there is much point in worrying about it (I should say, this isn’t a criticism of the paper I’m talking about, which took all the sensible measures to deal with the problems inherent in there analysis and presented their tree as something to work on rather than a final result). Similarly, a phylogeny that included data from everyone’s favourite animals the placozoans (Schierwater et al., 2009. doi: 10.1371/journal.pbio.1000020) was used to resurrect a hypothesis for the origin of animals that includes a flat ancestor, not unlike a modern placazoan or the ‘planula’ larvae of some cindarians. This scenario would again require nervous systems to evolve twice:

One of the reasons I’m skeptical of this idea comes from what we’ve learned about the genomes of simple animals. For my money, the discovery of Hox genes is the most remarkable result to come out of the the molecular revolution I talked about in the intro to this post. In turns out the genes that orchestrated your embryonic development have counterparts in Drosophila (or Sophophora if you’d rather) that do exactly the same job. If we move away from the bilataria, Hox genes are a bit harder to come by, but they still exist in the genomes of comb jellies, cnidarians and even palcozoans. But sponges have no Hox genes (Larroux et al, 2007. doi: 10.1016/j.cub.2007.03.008 and references therein). It’s conceivable that the sponges descend from an ancestor that had Hox genes, then gave them up, but it seems much more likely that Hox genes evolved after sponges parted ways with the rest of the animal kingdom.

The contents of animal genomes tell us more than which phylogenetic patterns are likely – they can also let us know how some of the unique features of animals might have arisen. Almost every time the genome sequence of a simple animal is published you’ll get a press release saying how surprising it is that genes associated with functions of more complex animals were found in this simple creature. And every time I read one of these press releases I sigh. Thesy are exactly analogous to someone that studies birds expressing surprise that therapod dinosaurs (the ancestors of birds) had forelimbs even though they couldn’t fly. You can’t make entirely new gene families out of nothing, new functions evolve when existing genes are retooled to attack different problems. The first animals faced two big problems when they moved from a single-celled lifestyle to a mulit-celled one: first, they needed to hold themselves together; and second, they needed ways for cells to communicate to their neighbours. Modern animals stick together with collagen (an entirely animalian invention) and a bunch of other sticky proteins including cadherins. The closest relatives of animals, the choanoflagellates which I talked about last week, have cadherins and other sticky proteins which they probably use to trap their prey (bacteria and other small cells).

Choanoflagellate genomes also give us a great insight into how the cells in the first animals started talking to each other. One of the most important pathways for cell-cell signalling in animals is called notch. Choanoflagellates don’t have notch genes, but they do have the raw material from which one could be made. The proteins that notch genes make have several distinct ‘domains’ each of which perform particular function – and all these domains are present in choanoflagellate genomes! These different domains could have been brought together by genetic recombination, creating new genes by so called ‘domain shuffling’. If we move just a little bit past the origin of animals and look at the genomes of sponges we can find the precursors of nervous system genes (see for example Liebeskinda et al., 2011. doi: 10.1073/pnas.1106363108).

So, that’s three posts about 5 000 words and a whole bunch of figures on the origin of animals. What can we say we’ve learned? Certainly, studying modern animals seems to be the only way to approach the question of where animals came from. The models we looked at last week allow us to understand how the problems associated with a switch from single celled life, to a colonial lifestyle and finally the emergence of an individual identity for the colony. Modern genomes also tell us how those changes might have happened at the molecular level. Sometimes, as is the case of the cadherins and the evolution of the nervous system, existing genes could be re-purposed to a new tasks. Other times, new features were likely achieved with the help of new genes, cobbled together by genetic recombination. We can probably disregard theories for the origin of animals that require complex morphological changes based entirely on molecular phylogenies, since those estimates are prone to biases and the most interesting branches in them remain uncertain. If we look at morphological similarities between choanoflagellates and sponges (which goes right down to the collared and flagellated cells they use to eat) it seems likely that the first animals were filter feeders that pumped water into their cells with long flagella.

If you forced to make a bet on what the first animals were like, I reckon they’d be a bit like a sponge larva. A collection cells similar to a choanoflagellate colony, moving about in the water column with their flagella. The big question is how that individual could have emerged from the colony, and here I like Paul Rainey’s ideas (described last week) about the way inter-cellular competition, which is bound to arise in a colony, could actually create the conditions that would foster cooperation. I’d be really interested to hear what any readers who made it all the way through this series think now.

Does a forty thousand year old finger point to another human species? David Winter Mar 26

No Comments

DNA extracted from a 40 000 year old finger bone found in a cave in Siberia might be evidence for a previously unrecognized human species. Or it might not be. The bone, which comes from what New Zealanders call a “little finger”, Americans call a”pinky” and paleo-anthropologists call the “distal manual phalanx of the fifth digit”, was found in the Denisova cave, in a region of Siberia from which remains of members of both our own species (Homo sapiens) and Neanderthals (H. neanderthalensis) have previously been found. The mitochondrial DNA (mtDNA) sequences generated from the finger bone are distinct from both modern human sequences and from previously published neanderthal sequences, but inferring species boundaries is a tricky business and the mtDNA sequences are not, in and of themselves, proof that the finger belonged to a member of a third human species.

Here’s the big figure from the paper, which was presented by Johannes Krause and colleagues in Nature yesterday. It’s a phylogenetic tree which relates the little finger’s mtDNA to H. sapiens and H. neanderthalensis sequences (click to see a high-resolution version):

The Denisnova sequence is red, Neanderthal sequences are in blue and modern humans are grey. So, the Denisova mtDNA forms a distinct lineage that isn’t represented in modern humans or in previously published Neanderthal sequences. By using the tree as the basis for molecular dating the researchers were able to estimate that Denisova lineage separated from other human mitochondrial lineages between 0.78 and 1.3 million years ago. The temporal context the molecular dating adds to the phylogenetic tree helps to us understand where this new mitochondrial lineage might fit into humanity’s family tree.

I’ve said before that most of our species’ history was played out in Africa, and, in fact, the same is true when we step up a taxonomic level and look at our genus. All the human species that have been found outside of Africa descend from migrants that moved out of that continent at some stage. Here’s a schematic representing some of the species in the wider human family tree and the timing of the migrations that moved them out of Africa.

How does the new evidence presented by Krausse et al. fit into that scheme? Perhaps the simplest interpretation is the the Denisova lineage represents a new species. The estimated age of the Denisova lineage makes it too young to have been carried out of Africa by the first wave of H. erectus migrants to leave Africa and apparently too old to have been inherited from the migrants that went on to form the Neanderthal lineage. If the Denisova sequence is something new then we’ll have to update our family tree, adding a new branch and a fourth migration out of Africa.

John Hawks thinks we should hold off on updating the family tree too qucikly. The Desinova specimen might be a Neanderthal. At first glance the tree presented by Krausse et al. seems to dispel that possibility since previously identified Neanderthal sequences are more closely related to modern human sequences than the new linaeage, but that tree is based entirely on mtDNA. The mitochondrial genome is inherited as if it was a single gene. We can often use trees estimated from a single gene (“gene trees”) as a proxy for species-level relationships (“species trees”) but, in fact, every gene in a population has its own history and there there are scenarios that can push a given gene tree away from underlying species tree. Perhaps the easiest way to visualise how you’d end up with mitochondrial lineages that diverged millions of years ago within a single species is to think about genetic lineages moving through a population while speciation happens. New species form when populations stop sharing genes with each other, in the diagram below the big black triangle represents a barrier to gene flow. What happens if multiple different gene lineages are present in the ancestral population at the time that this gene flow stops? Usually, given enough time, each species will “sort” into specific gene lineages that descend from just one of the lineages in the ancestral population, but it’s also possible for one (or both) species to maintain multiple lineages for some time. Such “incomplete lineage sorting” makes gene trees bad proxies for species trees and it’s just possible that something like this has happened in Neanderthals:

Perhaps by moving to the very Easterm edge of the Neanderthals range we’ve sampled for the first time a lineage that existed in that species for the whole time it was in Europe. Maybe, and Hawks surely knows a lot more about paleobiology than I do, but I don’t really buy it. It’s certainly possible for a species to harbour deeply divergent mitochondrial lineages, but the time it takes for gene-lineages to sort within a species is relative to the effect population size of that species. Neanderthals probably had a relatively small effective population size (and mtDNA definitely does, since only females pass it on and then in only one copy) making the retention of multiple lineages over hundreds of thousands of years seem like a long shot. As Hawkes argues, strong geographic structure in Neanderthal populations might have aided the retention of divergent genetic lineages against those odds, maybe the Denisova mitochondrial lineage was extinct in Western Europe but common in Central Asia? It’s possible, but I wouldn’t bet on it.

Finally, the Denisova sample might be our first look at H. erectus DNA. H. erectus remains have been recovered from China so it seems possible they were in Siberia too. As I’ve said, the molecular dating of the Denisova lineage probably makes it too young to be a descendant of the first wave of migration form Afirca (though, of course, there is some uncertainty associated with that dating), but it might be evidence of genetic exchange between African and the H. erectus diaspora. As we’ve come to understand the origin of our species we’ve realised that the simple “Out of Africa” model is just that, a model, and the true pattern is more complex. H. sapiens really did have its start in Africa and it really did push out into the rest of the world in the last 50 000 years or so, but during that expansion populations have continued to exchange genes. There’s no reason to believe that that H. erectus could not have done the same, perhaps the main thrust of the H. erectus expansion was 1.6-2 million years ago but genes continued to flow in and out of Africa for sometime after that.

So, there are three possibilities for the Denisova sample:

  1. It could be a new species,
  2. It could be an ancient mitochondrial lineage retained in eastern Neanderthal populations but lost elsewhere
  3. It could be the first H. erectus sequence.

We’ll need more genes (Krausse et al. report they are working on sequencing genes from the nuclear genome) or more complete specimens to know for sure but I’ll throw caution to the wind and say I think the first scenario to be the most likely and the second the least probable (remembering of course, that I’m not an anthropologist and these are pretty subjective estimates!). Perhaps I’m displaying some biases because I also think numbers one and three would be the cooler results. If either of those scenarios are true then we can add a third human species (alongside the Neanderthals and the ‘Hobbit’ H. floresiensis) that modern humans might have interacted with – it’s just so fascinating to imagine our ancestors living alongside other human species and how differently the world might have turned out if those other species had survived the few thousand years that separate us.

You should read Carl Zimmer’s post on the paper, he’s compiling expert opinions as they come to him. There’s also some more qualified comments via The Independent who made up for their poor news article on the story by having Chris Stringer from the Natural History Museum write a piece on it.

Krause, J., Fu, Q., Good, J., Viola, B., Shunkov, M., Derevianko, A., & Pääbo, S. (2010). The complete mitochondrial DNA genome of an unknown hominin from southern Siberia Nature DOI: 10.1038/nature08976

Did the Moa’s ancestor fly to New Zealand? David Winter Feb 04


New Zealanders often think of our unique biota as a sort of time capsule – a glimpse at lifeforms that have long since been extinguished in other parts of the world. New Zealand has been apart from the rest of the world for 85 million years. At that time the land that makes up our mini-continent split from the super-continent Gondwana, opening up the Tasman Sea and moving northward . A land apart from the rest of the world. Until recently most scientists have thought that the subset of the Gondwanan flora and fauna that set sail on that proto-New Zealand was likewise on its own evolutionary trajectory -insulated from biological happenings in the rest of the world. The idea of the New Zealand biota as a group of refugees from an ancient ecosystem hanging onto “Moa’s Ark” has become part of the New Zealand psyche.

In recent years Moa’s Ark has sprung more than a few leaks. Icons of our Gondwanan heritage like Nothofagus beech trees have been shown to be recent arrivals. Geologists have suggested the whole continent submerged 20 million years ago, drowning any refugees still on board, and now new research suggests even the most unlikely of immigrants – the giant, flightless moa – may have arrived in New Zealand well after we left Gondwana.

The group of birds to which the moa belonged, the ratites, have long fascinated evolutionary biologists. All the ratites are flightless (though, as we’ll see they are related to the quail-like tinamous which can fly passably well) and all the major landmasses that had their start in Gondwana had at least one ratite species before humans arrived on the scene. Africa has ostriches, South America the rhea, Australians have emus and cassowary, Madagascar had the Elepahant bird and New Zealand lost the moa but retains the kiwi. The far flung distribution of the ratites and their apparent lack of ability to disperse between continents has led to them being put forward as a classic example of an idea called vicariance biogeography in which the evolutionary history of a group is driven by the geological history of the land on which they live.

For vicariance biogeographers the evolution of the ratites was driven by the movement of the continents. The ancestor of all modern ratites was a flightless bird living in Gondwana and as each new continent split and rifted away from the super continent it took with it a population of ratites which adapted to the ecological changes brought on by their continent’s journey: cassowaries in the Wet Tropics of Australia, ostriches on the African Savannah, rhea on the Pampas. It’s certainly a nice story, but science has a way of ruining nice stories. The role of vicariance of evolution in the ratites was put to the test once we became able to use molecular evidence to reconstruct the relationships between species. If the geologically driven sketch of ratite evolution I presented above is right then the pattern of branching we find among ratites from different continents should match the order in which we know the continents broke up, something like this:

In 2001 Alan Cooper and colleagues sequenced the entire mitochondrial genome (some 12 000 bases of DNA) from representatives of each of the extant ratites and, remarkably, two species of moa. The long, careful process of retrieving DNA sequences from sub-fossil bones deserves a post of its own but for the sake of this article we only need to know what Cooper et al found when they used that DNA to recover the the relationships between ratite species.

The species in bold text above don’t fit the pattern that we’d expect from geology alone. If ratite relationships simply reflected the Gondwanan breakup we’d expect to see ostriches grouping with rheas (and apart from the other ratites). New Zealand’s two ratite orders are even more surprising, the kiwi lineage is more closely related to the Australian ratites than it is to the moa species. When combined with a molecular clock analysis Cooper et al. concluded that modern kiwis are the descendants of ancient immigrants hailing from either Australia or islands in the Lord Howe Rise (which have since submerged). In order to explain that trans-tasman dispersal the authors reached for the last resort of the desperate biogeographer and invoked a land bridge for which there is little geological evidence. In fact, as we’ll see it now seems more likely that the ancestors of the kiwi and the moa flew to New Zealand.

Even with the mitochondrial phylogeny of the group published there was considerable room for uncertainty in how the ratites related to each other. The underlying shape of ratite tree makes it particularly difficult to accurately recover with phylogenetic methods. When we use DNA sequences to estimate a phylogenetic tree we need to find species that share mutations that have accrued during the evolution of the group we’re looking at. The branches that relate the different ratite species are relatively short, so there was little time for mutations that set related groups apart from more distantly related ones to accrue. Even worse, the branches that reach to the modern species (the tips of the tree) are very long meaning there has been a lot of time to any mutations that did accrue in those critical short branches to be overwritten*. There are three approaches to dealing with this problem – sequence more genes (since each unlinked gene acts as a separate witness to the evolution of the group), sequence more samples (especially if doing so breaks up a long branch) or use a better model for the way mutations accrue in the genes you are studying. People have tried all three methods to get a better look at ratite evolution. Last year a group centred around the Field Museum in Chicago published a mutli-gene phylogeny of all birds that contained a big surprise for ratite evolution- the most recent common ancestor of all ratites flew.

As long as the ratites grouped together in a phylogeny it was reasonable to assume that they all inherited their flightlessness from the common ancestor of the group. The Field Museum study found that, in fact, the flying tinomous fit right in the middle of the flightless ratites. So, either the most recent common ancestor of the ratites and the tinamous flew and ratite lineages have subsequently lost that ability at least three times or that ancestor was grounded and the tinamous have rediscovered flight. In vertebrates the evolution of true flight has happened three times (in bats, pterosaurs and birds) while there are hundreds of examples of birds that have given up on flying. Moreover, a group of flying birds that are prone to flightlessness is hardly anything new – at least 30 species of rail (including our own weka and takahe) have taken to life on the ground. Given the ways the odds are stacked towards losing flight it seems probable the common ancestor that relates tinamous and ratites flew. The Field Museum study didn’t include any moa species and didn’t attempt any molecular dating so it’s hard to see just how the ancestors of the kiwi and the moa made it to New Zealand

A new study (I knew I’d get to it eventually) published in Systematic Biology throws some light on the New Zealand ratite story. Matt Philips and a team of researchers from the Alan Wilson Centre at Massey University took another look at the mitochondrial dataset used in Cooper et al’s 2001 study by adding more kiwi species and using models of DNA evolution that avoid some of the pitfalls of the ratite phylogeny’s difficult shape. The new ratite tree and a molecular clock analysis based on that tree confirm the idea of multiple loses of flight in the ratites and add a new finding – the closest living relatives of our giant moa are the quail-like tinamous:

So what does the new understanding of ratite relationships mean for our ideas about the origins of New Zealand’s ratites? The molecular clock doesn’t quite tick with the regularity of a stopwatch, so there is a good deal more uncertainty in the timing of the events presented above than the precisely defined nodes suggest. Still, even with the uncertainty of molecular dating taken on board we can safely say that both the New Zealand ratite lineages departed from their closest relatives after the Tasman Sea opened up.

The revelation that the tinamous are the moa’s closest living relatives suggests that the moa had ancestors that could fly. So, it seems the first proto-moa to arrive in New Zealand flew, or more likely was blown, here from Antarctica. Antarctica? It still seems amazing to me but 30 million years ago Antartica was still attached to South America and, without the circumpolar current to isolate from the world, was a relatively verdant continent. We know from fossils that Antarctica supported southern beech forests (still found in Chile, New Zealand and Australia) and marsupial mammals (strangely absent in New Zealand but still present in Australia and South America) so it’s no great stretch to propose the representatives of another Gondwanan group lived there. Antarctica certainly seems like a more likely jumping off point for dispersing proto-moa than South America, but either way it certainly seems they made it here under their own power.

The mode of dispersal for the kiwi’s ancestor is a little less clear. As we’ve seen we can be quite sure that they arrived in New Zealand after the Tasman Sea opened up and there is really no good evidence that there was ever a land-bridge across that sea. We can probably rule out walking. If we disregard the problem of dispersal for a second the simplest way to explain the distribution of flightlessness on the Phillips et al phylogeny is with a single loss of flight in an ancestor shared by kiwis and the Australian ratites. Under that scenario the kiwi would, presumably, have had to raft to New Zealand. Alternatively, given that we’ve seen the ratites seem to have an inbuilt propensity to becoming flightless we might imagine that the common ancestor shared by the kiwi and the Australian ratites could fly and each lineage has since lost that ability. In this case the kiwi could simply have flown from Australia to New Zealand (a journey that storms frequently inflict on Australian birds today). Without sufficiently old ratite fossils from either country it’s hard to choose one scenario over the other.Long range dispersal by rafting is probably an important force in biogeography but if I was forced to make a bet I’d put my money on ancient flying kiwis.

The radical rethink of ratite evolution that a decade of molecular phylogenetics has forced on us raises a lot of interesting questions. What it is it about the ratite body plan, development or behaviour that makes them so prone to flightlessness? Is that repeated loss of flight, and consequent lack of pressure to keep their weight down, enough to explain the trend towards gigantism? The authors of the most recent paper suggest both trends might be explained by ratites on each continent filling the ecological niches left by the extinction of the dinosaurs. The dates on their tree are certainly consistent with the idea that each ratite lineage independently took to the ground 65 million years ago but without more fossils and more precise dates for each split it’s very hard to test the idea further. I’m sure the story of ratite evolution has more surprises for us to uncover.


Outrageously geeky aside: these sorts of phylogenetic trees can even fall into the terrifying Felsenstein Zone in which the confidence with which you estimate the wrong tree increases as you throw more data at it.

Links to the primary literature are provided below but you should also check out Simon Collins excellent piece in the Herald and Mike Dickison, who got his PhD studying giant flightless birds and wrote about the idea that ratites flew to New Zealand way back in 2007.

Cooper A, Lalueza-Fox C, Anderson S, Rambaut A, Austin J, & Ward R (2001). Complete mitochondrial genome sequences of two extinct moas clarify ratite evolution. Nature, 409 (6821), 704-7 PMID: 11217857

Harshman, J., Braun, E., Braun, M., Huddleston, C., Bowie, R., Chojnowski, J., Hackett, S., Han, K., Kimball, R., Marks, B., Miglia, K., Moore, W., Reddy, S., Sheldon, F., Steadman, D., Steppan, S., Witt, C., & Yuri, T. (2008). Phylogenomic evidence for multiple losses of flight in ratite birds Proceedings of the National Academy of Sciences, 105 (36), 13462-13467 DOI: 10.1073/pnas.0803242105

Phillips, M., Gibb, G., Crimp, E., & Penny, D. (2009). Tinamous and Moa Flock Together: Mitochondrial Genome Sequence Analysis Reveals Independent Losses of Flight among Ratites Systematic Biology, 59 (1), 90-107 DOI: 10.1093/sysbio/syp079

Marsden Fund 2009: When Family Trees get convoluted David Winter Oct 08

No Comments

2009 has seen the 200th anniversary of Charles Darwin’s birth and the 150th birthday of the book that changed everything. 172 isn’t quite as round a number as the other two but as we celebrate Darwin’s bicentenary the we should remember July 1837. Sometime in that month Darwin opened his brown notebook, wrote the words “I think” on the top of a page and drew his very first phylogenetic tree. Over the next 10 or so pages (you can read them here) Darwin scrawled out the implications of the big idea contained in that tree – that both the diversity seen in modern species and the continuity of form seen between them might be explained by ancient species splitting to from distinct lineages, changing and splitting again. Life as a branching tree.

It says something about Darwin that having for the first time sketched out his idea he didn’t end with a hot blooded exclamation like Eureka! Instead he wrote cuidado – be careful. In fact, it took him 22 years of cautious letter writing and careful barnacle inspection (and one hell of a fright in a letter from Wallace) to publish The Origin. That book contained one illustration – a phylogenetic tree. In subsequent years the Tree of Life has become the central metaphor for Darwin’s view of life and recovering the relationships between organisms has become a major part of how we do evolutionary biology. New Zealand mathematicians have been among the leaders in providing biologists with the tools to produce phylogenetic trees and an understanding of how they might be best applied. This year the Marsden Fund has given grants to three projects looking at how to recover trees from the murkiest evolutionary histories.

Well resolved phylogenies help us to test evolutionary hypotheses – are New Zealand’s plants and animals relicts from Gondwana? Is Sphenodon guntheri really a distinct species of tuatara? Do kiwis descend from flying birds?* To put these sorts of questions to the test we need to estimate relationships between organisms, usually by reconstructing relationships between DNA sequences from them. In most cases these ”gene trees” are good proxies for the species trees we want to know about, but this biology so there are exceptions to the rule. Quite a few processes occur in populations and within genomes that drive gene trees away from species trees.

Charles Semple’s project “unravelling the web of life” attempts to account for the fact the as well as being passed from parent by offspring genes can jump from one species to another. Although we’ve known that such Horizontal Gene Transfer (HGT) is possible since the 60s it’s only more recently that it’s become increasingly clear that att the very base of the tree of life the degree of HGT was so great the tree metaphor with it’s simple one into two branching pattern breaks down, the tree becomes a tangled web. HGT is not limited to the base of the tree of life either, micro-organisms continue to swap genes with each other and your own genome contains quite a lot of virus DNA, inserted into your ancestor’s genome to get the ancient viruses multiply and passed on to you. Semple’s project aims to provide the same sort of mathematical basis we have for understanding tree like evolution for the much more complex pattern we see in evolutionary histories that include HGT.

James Degnan‘s project moves from the root of the tree of life to the tips. When the the process that forms new branches on trees, speciation kicks of each of the nascent species will inherit a suite of genes, each of which has its own history (the same lineages I’ve written about with respect to my mitochondria). If we try and infer species level relationships from just one gene’s tree we fall into all sorts of problems, especially if we are dealing with recent or rapid speciation.(there are even genes that would place Gorillas as humans closest living relatives). Degnan has already worked on one of a bunch of programs avaliable to infer species trees from multiple gene trees and his “fast start” Marsden grant will allow him to keep working this field.

It’s not just strange genetic quirks that strain Darwin’s metaphor of tree like evolution. In plants, and at least some animals speciation sometimes occurs following hybridisation – two barnches into one. Even hybridisation isn’t making new species it’s a force that can provide conflicting phylogenetic signals to sequence data . Barbara Holland‘s grant for “Untangling complex evolution: when the Tree of Life is not a tree at all” will help to make better methods for revealing these kind of complex evolutionary histories and help biologists know when recovering the One True Tree for a groups isn’t a sensible goal.

*The answers are “mostly not”, “no” and “almost definitely”

Network-wide options by YD - Freelance Wordpress Developer