Wednesday, April 24, 2013

Genome-wide recombination patterns and their implications in threespine stickleback fish


[This post is by Marius Rösti; I am just putting it up.  -B.]

During meiosis – the characteristic cell division in sexually reproducing organisms – the mother cell that gives rise to sperm or ova has to multiply and reduce the number of chromosomes from two full sets to one. During this process, the two homologous chromosomes (one from each parent) can exchange DNA segments. This process is called meiotic recombination. In a paper just published in Molecular Ecology, we demonstrate in one of the prime model systems in evolutionary biology (threespine stickleback fish) that recombination is distributed highly heterogeneously within the genome, and highlight implications for genome evolution and its empirical investigation.

A threespine stickleback, Gasterosteus aculeatus

Heterogeneous recombination rate

Threespine stickleback (Gasterosteus aculeatus) offer a powerful model system in evolutionary biology. Some of the reasons are the well-known natural history and ecology of this fish, as well as the availability of many genetic and genomic tools. However, a detailed analysis of recombination in stickleback has been missing so far. To do this, we first reassembled and improved the stickleback’s reference genome. Based on 300 individuals and approximately 2000 genome-wide SNP markers, we find that the recombination rate is highly elevated in the chromosome peripheries relative to the chromosome centers. A similar distribution of recombination events along chromosomes has recently been found in several other taxa, including humans. We further detected a minimum of one recombination event per chromosome (but not chromosome arm) per meiosis event. These findings likely point to strong functional constraints on the rate and distribution of recombination within the genome.

Heterogeneous recombination rate drives patterns of genetic diversity and population divergence

Incorporating genome-wide sequence data from four natural stickleback populations inhabiting ecologically different lake and stream habitats, we can demonstrate clear associations between recombination rate and the magnitude of allele frequency shifts between populations, and between recombination rate and genetic diversity within populations. In these young (only a few thousand years old) populations experiencing divergent natural selection, these patterns certainly reflect genome-wide heterogeneity in the effect of selection on linked sites, which has proved hard to demonstrate convincingly in non-ecological (genetic) model systems such as Drosophila flies. This recombination-driven heterogeneity in signatures of selection has a potentially important methodological implication: ecological genome scans will detect divergence outliers more easily in low-recombination regions. Furthermore, we detected a strong association between recombination rate and GC nucleotide content. As suggested in other organisms, this pattern perhaps arises from GC-biased gene conversion, potentially reflecting a direct influence of recombination on genome evolution.

Recombination and sex chromosome evolution

As in humans, in threespine stickleback  males are the heterogametic sex: males carry an X and a Y sex chromosome, while females carry two X’s. We were able to confirm and narrow down the physical boundaries of a previously inferred small ‘pseudoautosomal’ region within the sex chromosome where recombination between the X and Y still occurs. The rest of the X chromosome does not recombine with the Y any more, but it does exhibit two regions characterized by distinct levels of differentiation between the X and the Y. Such ‘evolutionary strata’ of Y-degeneration are expected when the suppression of recombination between the X and Y occurred in discrete pulses across large chromosomal regions. This first demonstration of evolutionary strata of Y-degeneration in a fish species highlights the devastating effects of suppressed recombination during the evolution of sex chromosomes.

Paper reference

Roesti, M., Moser, D. and Berner, D. (2013). Recombination in the threespine stickleback genome—patterns and consequences. Molecular Ecology. doi: 10.1111/mec.12322

Saturday, April 20, 2013

Of mites and men

[This post is by Thomas Cameron; I am just putting it up.  -B. ]


Those that observe ecological dynamics have always stated that evolution plays a role, but it is only relatively recently that there has been a fever of interest in quantifying what that role might be. It is not the case that there was ever a lack of interest; it was just generally assumed that as evolution occurs on very long timescales, it will have little influence on contemporary ecological dynamics – certainly on the timescales over which ecologists would want to make predictions. But increasingly we have come to accept that phenotypic change can occur very quickly, in only a few generations, and we have a good body of evidence showing that “rapid” or “contemporary” evolutionary change in the underlying biology of species is common. (As an aside, I wonder if it should not be called “rapid” evolution, as it has recently been pointed out to me that palaeontologists see little more phenotypic change within species in thousands of generations than ecologists see in tens of generations.)

Much of the recent interest in contemporary evolutionary change was generated by the theory of harvest-induced selection, a framework in which it is thought that high trait-selective harvesting rates lead to a change in the frequency of traits in a population, diminishing the type of animal or plant that the harvester is seeking. More recently further interest has come from the question of species responses to current rates of environmental change. And so it was with these questions in mind that a project was born. What exactly is the contribution of evolution to the way in which populations respond to their environment, and how is this influenced by high (but reasonable) harvesting rates?

Our project was to compare the adaptation of populations of soil mites, taken from the wild and placed into environments differing in the form of environmental variation caused by experimental manipulation in the variance, but not the mean, of the food they received. In a recent paper (DOI: 10.1111/ele.12107) I presented the results of one of these environmental treatments: periodically varying food. The variation of the food is not so important for this blog post, so I won’t discuss it again here. Overlaid upon this, we harvested 40% of either juvenile or adult individuals each week, with an unharvested control. This harvest rate was estimated to be close to a maximum sustainable yield, based on harvesting adults. Other important information is that these soil mites have essentially three demographic stages: eggs, juveniles and adults. They are sexually dimorphic as adults and have a generation time of approximately five weeks.



A mite.  Photo credit: Dr. Tom Cameron.


For two years we counted the numbers of the different stages of soil mites each week in each of 6 replicate population tubes per harvesting treatment (18 tubes). We combined this census of demography with a census of the life history of the mites. For the uninitiated, a life history is a way to describe the balance between how an organism invests its available resources into survival, growth or reproduction. To conduct the life history census, approximately every 15–20 weeks we took a sample of mites from two of the population tubes per treatment and reared them in a common garden for three generations (a common garden is where you rear all organisms in the same conditions, which helps to minimise any non-genetic effects of the parents’ environment on the life history of the mites). Our assumption is that if we view the life history of the mites after they have been in the common garden then any differences that remain should be mostly genetically based. Because that is an assumption, we also conducted a third census of the genetic dynamics in those two sample populations per treatment so we could estimate whether any changes we saw in the life history of the mites was “likely” to be caused by natural selection.

We have two main results. First, by placing mites from the wild into closed populations in the laboratory we created populations not suited to their new environment and the populations began to decline toward extinction. However, after only five generations this decline slowed and then reversed into a long-term increase. This dynamic occurred in all stages in each of the 18 tubes across all harvesting treatments. Our analysis of the life history found that there was a large delay in mite growth rates occurring during this period. Normally we think of a delay in growth to maturity as a negative response to poor conditions, but under the conditions the mites found themselves in compared to their wild ancestry, it was selected for. To understand why, we also looked at survival and fecundity. While survival was little changed, those mites that delayed their growth to maturity and spent a longer time in the juvenile stage could double their peak reproductive output. To confirm this pretty adaptive story that mites investing in increased individual reproductive output could rescue their population from extinction, we compared the population growth we would predict based on the life history of an average individual with the real population growth rates we observed in the population time series, and they were highly correlated. Indeed, if we do not use the changing life history trait values we observed throughout the experiment we cannot predict the recovery and long-term population size increase we saw in the experiment. Finally, we were able to show that this delay in age-to-maturity was linked to patterns of change in the genetic diversity in the populations, suggesting that natural selection was an important factor in the ecological dynamics even over just a few generations.

Harvesting had both an ecological and an eco-evolutionary effect on the population dynamics. Its eco-evolutionary effect was to prevent individuals from juvenile-harvested populations from reaching the optimal phenotype, as they could not stay in the juvenile stage too long or the populations would go extinct. Adult-harvested populations produced individuals that delayed maturity a little longer but matured larger, which we suggest is an adaptive response that increases fecundity early in adult life. These results are exactly what we would predict from current theory. Certainly the ecological effects of harvesting were strong, in that when we ceased harvesting the population dynamics of all treatments were more similar; but differences remained for many generations, especially in terms of population variation.

So our take-home message from our first result from this project is that ecology and evolution are not as separate as we have traditionally assumed. Even over a few generations we can see that evolution has a role to play in how populations respond to environmental change, whether biotic (e.g. harvesting) or abiotic (e.g. climate change).

There has been some criticism of microcosm studies of harvest-induced evolution that they produce trait change underpinned by selection far faster (i.e. 10% per generation) than the slow rates observed in the heavily exploited wild populations (1–2%). In our study, in which mites were simply enclosed in population tubes and exposed to strong density-dependent competition for food and then harvested near the maximum sustainable yield, we saw a 4% per generation response to environmental change and a 1.4% per generation change in response to harvesting. This clearly places the trait changes we saw in the “slow” category and suggests that even slow rates of harvest-induced evolution can have significant and long-term effects on population dynamics. Our results are therefore further evidence that evolutionary considerations should be taken into account in harvesting management. By this I do not necessarily mean the harvest rate, as I think ecological models can mostly deal with this; rather, I mean in terms of making realistic predictions regarding whether harvested populations will recover from overexploitation, and how long such recovery might take.

How important are these results of contemporary/rapid evolution? Well, the truth is we do not know. Many studies have concluded that rapid or contemporary evolution may have repercussions beyond population ecology, for example in ecosystem services and ecological community structure, resilience and function; but to date there are few (if any?) examples in which this has been demonstrated. For this reason, I would suggest that future studies of the role of evolution in ecological dynamics should turn toward community ecology.

Tuesday, April 16, 2013

CEEB 2013


Every year, the Conservation, Ecology, Evolution, and Behaviour group (CEEB) of the Department of Biology at McGill organizes a retreat for faculty and graduate students. When I say retreat I mean a cocktail of science talks, workshops, trivia, lots of wine and even more fun, with a final touch of Kommando Pimperle. As usual, the event was held in the beautiful Maison Gault in Mont Saint-Hilaire. It was the first time that I was able to attend one of these, so I thought I would write a short post about it.

The retreat began with a series of ignite talks, which are basically 5 min presentations where the slides are automatically changed every 30 seconds. This is the first time that this format was used at the retreat, so it was interesting to see how different people dealt with the pressure of time and the unexpected change of slides. Ignite talks definitely have some advantages, particularly because most people are able to keep their ADD on a leash for 5 min. Moreover, you don’t have time to spare on irrelevant stuff; you go straight to the main point. But things can get awfully complicated if you are not prepared, or if your topic is too complicated. Some of the talks were really interesting, but they were so fast that I already forgot the details. However, I do remember that topics ranged from NSERC funding, Fowler’s toads, evolution of cognition, smaller and smaller R2s, ecoevolutionary dynamics, flowering time and climate change, ant evolution, among many others.

Jonathan Davies and Andrew Hendry gave a workshop on how to write research papers, and how to deal with reviewers if your paper is rejected. I really don’t know if the latter was added because it was assumed that if you did as suggested in the first part you were likely to be rejected, but I hope not. Monica Granados and Eric Pedersen gave another workshop, but this one focusing on how to make sexy plots using R and Adobe Illustrator. If you attended this workshop you were also less likely to worry about the reviewers mentioned in the first one.

I think this is some sort of water polo on chairs and without the water, bathing suits, and all of that. 


After dinner it was time for trivia, but not just any trivia, it was a CEEB trivia. Questions were related to the talks and general knowledge in Biology. I won’t go into detail here because I spent most of the time trying to get the other teams to answer wrong, rather than actually helping my team  –a good reason for which we didn’t win, sorry les tetines.


Some of the post trivia survivors



The night ended –very late– with a German game called Kommando Pimperle. I want to be honest here and say that until I started writing this blog I thought the game was Kommando Bimpula, not even google knew what I was talking about. Eventually I found out what the real name of the game was, and it turns out that it is quite popular! It is also very fun to play. Basically, there are several instructions for movements with your hands that are given by the person leading the game, but they can not be performed by the rest unless the word kommando is given before. It gets even trickier! The person leading can also trick you by combining different movements and throw the kommando word wherever s/he likes. And to make it even worse, it is all in German! I was also usually the first person to lose…

This is a video of Kommando Pimperle at a cafeteria in a German university.


All in all, my first CEEB was a great experience! I got to know (in 5 min) what a lot of the people in the department are doing, and got to talk about science in a more informal environment. It is a good way to interact with your peers and establish future collaborations. It could also be the first step to learning German! Kommando Pimperle…


Sunday, April 7, 2013

Evolutionary Rescue, Pappy Van Winkle, and Next-Level Smokin-Hot Secret Sauce


Many years ago, when Mike Kinnison and I were office mates in Seattle, we started a list of “words we should use more often.” These were usually esoteric English words that we had encountered in some publication and decided were just too cool to be used so infrequently. I can’t remember all of the words but consanguineous was certainly one of them. I did use consanguineous in at least one paper but, sadly, it did not precipitate a particularly far-reaching or long-lasting meme. But what if I had tried harder? What if I had insisted that my fellow graduate students use the word in their papers – or bribed them to do so? What if I had used it in all of my own papers? What if I had extorted (or bribed) each visiting seminar speaker to use it in their talks? This is precisely the experiment currently being conducted by the Jen Schweitzer and Joe Bailey labs at the University of Tennessee.

While visiting UT for a seminar last week, I met with Joe and Jen’s students. At the end of our meeting, the students casually mentioned that they had come up with a series of phrases that should be introduced into the scientific lexicon – and they pointed out that I could help their cause by using the phrases in my seminar later that day. I immediately thought back to consanguineous and its ignominious continuance in anonymity. Maybe I here had a new chance to save some cool lost word or phrase from the dustbin of academia. I would be glad to help, I told them, what are the phrases? They pointed to the chalkboard behind me, where I read:

“Next level shit”
“Smokin’ hot right now”
“There ain’t no secret sauce”

Hmmm – not quite what I was expecting and perhaps not so deserving as consanguineous but, then again, who am I to quash enthusiasm and ambition. The meme does not stop here. After presenting some genetic data in my talk, I pointed out that the particular genetic markers (microsatellites) I used were rather old school, and that what we really needed to answer the question was some “next level shit.” I then pointed out that what is “smoking hot right now” (actually I had forgotten the phrase and needed some prompting) is RAD-tag based SNP discovery – for which I conveniently had some results in my next slide. Two phrases down, one to go: the hardest one. I struggled to think of an appropriate use for “there ain’t no secret sauce” and eventually realized, while looking at my conclusions, that this was precisely the spot.

Now, I can’t say that I will continue to use these phrases in all my talks, but I do feel I have done my part and that, should the meme not take off, I will at least have given it the "old college try.” If it does take off, I suggest that consanguineous should be next. In fact, I anticipate that all of Jen and Joe’s students will now feel obliged to use it in their talks. (It means “of the same blood” and so can be used in relation to ancestry or relatedness.)

Jen and Joe holding court in the Smokies.

During my visit, I stayed with Jen and Joe, who kept me well fed (home-made crab cakes, gumbo, and cherry pie), well beveraged (beer, wine, and – in a rare treat – a drink of 23-year-old Pappy Van Winkle’s Family Reserve Bourbon), and well entertained (Emmylou Harris, Rodney Crowell, and Richard Thompson at the historic Tennessee Theater). Joe even let me beat him at ping pong on his home court. I stayed through Saturday to see some sights in the nearby Smoky Mountains, a major biodiversity hotspot. Joe and Jen both had field courses to teach in Cade’s Cove – a mountain valley in the Smokies – and so I tagged along. Both courses – one graduate and one undergraduate – were to collect data at a series of deer exclosures. 

Deer exclosure - or Hendry enclosure?

At one site, Mark Genung and Joe showed me a dead hemlock tree and then found a hemlock sapling that was covered with an invasive scale insect. This insect has apparently decimated hemlock populations on a massive scale. This brought to mind many seemingly parallel instances: Dutch elm disease, myxomytosis in Australian rabbits, phyloxera in European grape varietals, sudden oak death, white nose syndrome in bats, tuberculosis in native Americans, the Black Death in Europe, mountain pine beetles, chestnut blight, and chytrid fungus in amphibians. In each case, an emerging disease – often (maybe always) an invasive, or at least spreading, species – decimates native populations that are not resistant.

What I find interesting about these catastrophes is that they rarely cause species extinctions – except perhaps for chytrid fungus. Instead, the massive declines are arrested short of extinction and the native species either carry on at a much lower abundance or ultimately recover. The interesting question for me is why extinction does not occur. Three possibilities come to mind. First, success of the disease may be frequency dependent, such that its impact or ability to spread greatly decreases as the host becomes rare. This makes some sense as the spread of a disease often depends on the number of nearby susceptible hosts – and so a decline in population density of hosts will decrease the chance that the remaining individuals will be infected. Second, hosts may evolve resistance – as long as genetic variation in resistance exists, then the individuals that survive and reproduce will increase the frequency of resistant genes. In fact, massive mortality events are expected to drive the fastest rates of evolution – because they can impose the strongest selection. Third, the disease may evolve to be less severe, as would befit its continued existence. I have no idea which of these effects is most important in any of the above examples, but it seems to me an important eco-evolutionary question in the context of evolutionary rescue.

Evolutionary rescue is the idea that when environmental change results in maladaptation that causes a population decline, adaptive evolution might reduce maladaptation and thereby arrest the population decline and allow recovery. Evolutionary rescue is generally thought to be most effective for organisms with short life spans, such as bacteria, viruses, or some weeds and insect pests. This makes good sense because these short-lived and numerous organisms presumably have high genetic variation and mutation rates and thus greater evolutionary potential. But it seems to me that large and long-lived organisms, such as trees, have something else going for them. In particular, high mortality can eliminate all but the few mature individuals that are most resistant, which – owing to their very high reproductive output (a birch tree can produce 15-17 million seeds per year) – have the potential to rapidly recover population size. I am not saying that bacteria and viruses don’t have the advantage in evolutionary rescue, merely that the supposed disadvantage to long-lived organisms might sometimes be partly offset by a combination of extremely strong selection and high potential reproductive output in survivors. For more about evolutionary rescue, see the recent PTRSB special issue.

A Tennessee Turkey strutting its stuff.
Well, that’s it for now. I hope to enjoy the rest of my day in the airport of some city that ends in “ville” (Tennessee has more than 50 such cities and towns – more than any other state) and some city that ends in “ark” before finally making it home to that city that ends in “real”.

In the Newark Airport, is it the Earl himself? Or maybe a descendent, perhaps basking in the glory of his ancestor - or protesting the lack of royalties.


Tuesday, April 2, 2013

Carnival of Evolution #58

Carnival of Evolution #58 is now up!  Although they got the name of our blog wrong, Felipe Dargent's post on mixed-species groups made it in!  A month later than it was supposed to go in, but nevertheless, it made it.  :-O  Joost Raeymaker's post on parasite-induced evolution in cichlids was supposed to be our nomination for this month, but, well, there has been a bit of confusion lately.  Perhaps it will make it into next month's Carnival.  :->  Check out Synthetic Daisies to see all the other groovy evolution-related posts up there now!

The theme this month is the future of evolution, so here's a possible future for human evolution.  Dig the threads, man.


The Hendry Lab in the year 2380.

A 25-year quest for the Holy Grail of evolutionary biology

When I started my postdoc in 1998, I think it is safe to say that the Holy Grail (or maybe Rosetta Stone) for many evolutionary biologists w...