Top

iBiology

Bringing the World's Best Biology to You

Genetic Variation and Phylogeography

Part 1: Gene Trees and Phylogeography

00:00:00.15 Hi, my name is Scott Edwards
00:00:02.06 and I'm a professor of organismic and evolutionary biology at Harvard University.
00:00:06.22 And today we're going to learn about gene trees and phylogeography.
00:00:11.21 Phylogeography is a very exciting field combining genetics and ecology
00:00:19.06 and increasingly, information from geographic information systems.
00:00:24.23 And so we're going to learn how to combine genetic variation with these other fields
00:00:29.21 how to measure genetic variation and how to use the patterns in the gene trees
00:00:35.22 to infer something about population history.
00:00:37.29 So, what is phylogeography? We're going to first talk about that
00:00:44.22 and then we're going to talk about how do we measure genetic variation
00:00:49.03 within a species. This is really important because the amount of genetic variation within a species
00:00:54.22 can tell us a lot about how healthy that species might be,
00:00:59.03 whether it has the possibility of adapting to environmental change such as climate change.
00:01:05.25 And then, we're going to talk about an interesting way of looking at genetics diversity
00:01:11.13 and that is through so called gene trees.
00:01:14.28 As we'll see, gene trees are a very intuitive and natural way to look at genetic variation within a species.
00:01:22.19 And then finally, we're going to talk about how do we look at natural selection.
00:01:28.06 How can we tell whether genetic variation is not being influenced by just neutral processes,
00:01:36.04 just random processes of population genetics but might actually be influenced by
00:01:42.19 deterministic processes like natural selection.
00:01:46.08 So, to begin, phylogeography is using genetic variation
00:01:54.12 and looking at it in the context of gene trees or genealogies of alleles
00:02:01.06 and linking those genealogies with the geography of the genetic variation.
00:02:06.23 As its name implies, phylogeography combines a little bit of phylogenetics,
00:02:13.25 looking at the relationships of alleles and genes to each other,
00:02:17.22 as well as geography.
00:02:20.04 How do we lay those lineages on a map and infer something about the history of populations?
00:02:26.23 So let's dive right in and talk about some of the history of phylogeography
00:02:32.04 and why it's such as special field today.
00:02:34.03
00:02:38.00 Looking at genetic variation within species is an old pastime.
00:02:42.09 It's answered a lot of riddles about the nature of species themselves.
00:02:47.10 And this slide here shows a picture of a type of data called allozyme electrophoresis.
00:02:56.26 It was a very common tool used in the 1960s and early 70s
00:03:02.21 to look at the amount of genetic variation within species.
00:03:06.02 What you can see here are bands on a gel.
00:03:10.22 You can see that the bands vary in their distance from the top to the bottom.
00:03:15.11 And what these are are protein variants.
00:03:19.16 Each lane, going from top to bottom, represents a different individual
00:03:24.02 and the different bands you can see are the different alleles
00:03:27.21 within one particular locus, one particular gene,
00:03:30.23 in this case, in the fruit fly Drosophila pseudoobscura.
00:03:35.25 What you can see is that there is a lot of genetic variation;
00:03:39.14 different individuals look different.
00:03:42.05 These alleles are being separated by the charge, not the size of the molecule, but the charge;
00:03:50.11 whether it's a negatively charged allele or a positively charged allele.
00:03:53.19 And you can see there's a lot of genetic variation.
00:03:56.15 And this was scientists' first view of the amount and nature of genetic variation within species.
00:04:03.10 It was a very exciting time and as you can see this paper was published in 1966.
00:04:07.20 Let's fast forward now to a different approach
00:04:13.13 which was introduced to the study of natural populations in the late 1970s.
00:04:19.19 And one of the pioneers in this field was a man named John Avise
00:04:25.07 and this image is from a book that he published in the mid-90s summarizing these results.
00:04:30.29 This new technique was called restriction enzyme analysis.
00:04:35.13 It used what molecular biologists call restriction enzymes to cut DNA into fragments.
00:04:44.08 Each restriction enzyme (these are proteins that come from bacteria)
00:04:49.14 comes from a specific species of bacteria and cuts the DNA at a specific recognition site
00:04:58.07 which is indicated by anywhere from 4 to usually 6 letters in the DNA sequence.
00:05:05.10 Every time that protein sees that sequence of letters, it will cut the DNA at that site.
00:05:11.13 And what you can see again on this gel
00:05:15.14 each lane going up and down represents a different individual.
00:05:19.25 In this case, we're using the restriction enzyme called EcoRI
00:05:24.19 and that tells us it's from E. coli.
00:05:26.20 You can also see on this gel, there's a lot of genetic variation between individuals.
00:05:32.29 You can see that the different lanes look different.
00:05:35.20 So, for example, lanes D, C, and E, you can see there's a different banding pattern
00:05:42.15 than the lanes A and B, those group of lanes A and B.
00:05:48.01 And what that means is that lanes D, C and E have an extra sequence in their genome
00:05:55.04 that allows this protein to cut there.
00:05:57.26 It's a way of detecting genetic variation.
00:06:00.17 And this new way of detecting variation was quite different from
00:06:04.27 the previous way I showed you, allozyme electrophoresis.
00:06:08.06 And critically, as you can see on the bottom of this slide,
00:06:11.08 by cutting the DNA in particular places,
00:06:14.15 we can begin to form ideas about how the different variants are related to each other.
00:06:20.21 They're not simply alleles in the population that are different from each other
00:06:27.01 and we can score them as the same or different.
00:06:29.17 We can actually begin to related those alleles to each other in a simple genealogy.
00:06:35.15 And that's what you see below the slide.
00:06:39.00 You can see a network of connections between the different so-called haplotypes
00:06:44.11 indicated in each lane of this gel.
00:06:47.02
00:06:49.22 What we'e done is we've moved from what could call the older population genetics
00:06:54.19 which was based primarily on what we call allele frequencies.
00:06:59.08 We could count the number of individuals that have allele A in a population at maybe 40%
00:07:06.17 and the number of individuals that have allele B which may be 60%.
00:07:10.24 And for a long time, that's how population geneticists and phylogeographers
00:07:16.27 would make inferences about the population history.
00:07:19.14 Allele A would be present in 40% in one population, 60% in another population
00:07:25.24 and from that we could make inferences about what had happened to those populations.
00:07:30.11 That was the old way of doing things.
00:07:32.20 The new way, which I think is much more exciting
00:07:35.27 takes into account the genealogy of the alleles themselves.
00:07:41.10 Instead of just counting the frequencies of alleles in populations
00:07:45.22 we can actually relate those alleles to each other in a genealogical tree
00:07:51.08 or in a phylogenetic tree.
00:07:53.05 So the new view...of course we can calculate the frequencies if we want,
00:07:58.11 but we have this additional power of connecting the alleles together in a genealogical tree,
00:08:04.02 and as you'll see, this allows us to really infer some interesting things about population history.
00:08:09.03
00:08:11.15 Also, an interesting difference from the earlier population genetics
00:08:15.08 is that whereas the older population genetic approaches
00:08:18.07 tended to think of genetic variation going forward in time,
00:08:22.09 We would simulate what gene frequencies would look like
00:08:26.02 from an ancestral population to a more recent population.
00:08:30.05 In this new approach, which is actually called coalescent theory,
00:08:34.23 because the alleles are coalescing into common ancestors.
00:08:38.24 In this new approach, we're actually going to think of it going backwards in time.
00:08:44.05 We can actually make inferences from just the sample of alleles that we have collected.
00:08:49.06 And that greatly simplifies thinking about phylogeography and population genetics.
00:08:55.27 We don't have to worry about the entire population,
00:08:58.18 just the sample that we have looked at in our particular study.
00:09:02.23
00:09:07.10 OK, so here's an example of a gene tree.
00:09:11.00 This is, in fact, the first gene tree; published in the late 1970s by John Avise and colleagues.
00:09:17.24 It relates to alleles in a group of rodents called deer mice.
00:09:23.17 You can see Paramiscus polionotus, Paramiscus maniculatus, and Paramiscus leucopus.
00:09:30.01 And the particular gene we're looking at is a very important molecule called mitochondrial DNA.
00:09:39.05 And this was a really remarkable paper
00:09:43.15 because it showed how you could relate different alleles to each other in a genealogical tree.
00:09:50.06 We won't talk about how we actually construct those relationships.
00:09:54.25 That's another field called phylogenetics.
00:09:58.08 Take it for now that we can assemble the different genetic variants into a genealogy.
00:10:03.15 And what you can see here, for example, is that the alleles within maniculatus
00:10:09.01 are most closely related to the alleles in polionotus.
00:10:12.11 They share a common ancestor.
00:10:14.16 Outside of that we have additional alleles from Paramiscus leucopus
00:10:19.28 which is the most distantly related species in this group.
00:10:23.19 This opened up a whole new way of looking at genetic variation
00:10:27.04 and was really quite exciting for starting off the field of phylogeography.
00:10:32.01
00:10:35.17 Now, an important concept to think about in much of subsequent discussion
00:10:40.02 is this idea of effective population size.
00:10:43.15 It's a very important term in population genetics and in phylogeography.
00:10:48.11 It's usually abbreviated as a capital N or a capital N sub e.
00:10:54.06 What is effective population size?
00:10:57.12 The effective population size of a species is the size of an ideal population
00:11:03.25 whose dynamics, whose evolutionary and population dynamics,
00:11:09.12 mimic the actually population that we're studying.
00:11:13.19 The assumption here is that no population in nature is simple.
00:11:18.22 There's a lot of complicated dynamics going on.
00:11:21.05 Populations are going extinct; being re-colonized by new individuals.
00:11:25.18 Populations are experiencing flow of genes into them and out of them.
00:11:31.04 Populations are changing in size constantly.
00:11:34.19 All of these are very complicated dynamics.
00:11:37.19 What we want to do is to summarize a basic characteristic
00:11:42.06 of those real populations in a single variable
00:11:45.28 and that variable is effective population size.
00:11:49.01 It's the size of a very simple population that has the same dynamics
00:11:55.02 as the very complex population that we're studying in nature.
00:11:58.03 So, by dynamics I mean things like how genetic diversity there is in a population,
00:12:03.20 what's the rate of loss of genetic variation,
00:12:06.20 if that population becomes isolated from other populations
00:12:10.08 or it can be something like the change in allele frequencies over time of those real populations.
00:12:17.07 By ideal, I mean a single population where there's no natural selection at all.
00:12:27.04 Everything is neutral
00:12:28.06 meaning allele A and allele B don't change the fitness of the individuals' bearing those alleles.
00:12:35.05 Every allele is, in terms of fitness, the same.
00:12:39.15 There's no population structure
00:12:42.16 meaning there's no subdivision of this population into more fine units.
00:12:47.29 And we have completely random mating within that single population.
00:12:52.12 So, it might be surprising that we can actually encapsulate
00:12:55.20 all the complexities of a real species into a single variable
00:12:59.12 and in fact, there are a lot of assumptions that go into this exercise.
00:13:03.00 But, still, the effective population size is a very important way of describing populations.
00:13:09.22 Let's look at some examples of effective population size
00:13:15.00 and how they differ from real populations.
00:13:18.04 Here's an example of an effective population size which is based on the sex ratio in the population.
00:13:24.11 For example, say you have an elephant seal and we know of course that elephant seals
00:13:29.08 typically have one or a few males that mate with all of the females in the population.
00:13:35.29 So, you may have 25 individuals total, with 1 of them being a male, 24 of them being female.
00:13:45.13 So our census size in that case is 25, the actual number of individuals.
00:13:51.05 However, the effective population size of that species
00:13:55.19 one can calculate by the equation that you see,
00:13:58.15 which is simply 4 times the number of males, which is 1, times the number of females, which is 24.
00:14:05.11 So that's 4 times 24 which is 96.
00:14:09.02 Divide that now by the sum of the number males and females which is 25.
00:14:14.21 And what you're going to get is a number approximately 4.
00:14:19.06 And now you can see how the effective population size is much, much smaller
00:14:24.25 than the census size which is 25.
00:14:27.20 And that's because there's this deviation from equal sex ratios.
00:14:32.15 In this population we've got 24 females and a single male.
00:14:36.18
00:14:39.13 Let's look at another example of effective population size.
00:14:41.28 Here's a case where we have an isolated population.
00:14:45.05 It's not getting any new genetic diversity from other populations.
00:14:49.05 And it's also not getting any new diversity through new mutations.
00:14:54.10 Let's talk about what's going to happen to genetic diversity in that population.
00:14:58.29 It's going to decline over time just by chance.
00:15:02.27 Alleles will get lost. Individuals with those alleles won't breed.
00:15:06.11 And you can see here that our measure of diversity, which is H,
00:15:11.11 how that declines with time, which is the little t (time is measured in generations),
00:15:18.24 and, you can see our friend N, our effective population size.
00:15:23.15 We can see how we would predict the loss of heterozygosity over time
00:15:28.24 with each generation if this population were an ideal population of a certain size.
00:15:34.17 And that's what you can see on the graph here
00:15:36.22 and that is an example of effective population size based not on males and females,
00:15:44.22 but now based on the loss of genetic diversity over time.
00:15:48.10 Another example is what we call the variance effective population size.
00:15:56.02 In this case, the variance in the number of matings of say males
00:16:02.03 will cause the effective size to shift from the census size.
00:16:07.23 This effective size is similar to the sex ratio effective population size that we saw 2 slides ago.
00:16:14.08 In this effective population size, N is now our number of breeding individuals in the population.
00:16:23.23 Could be our census size.
00:16:26.13 V sub k is the variance in the number of matings, for example, per individual or per male.
00:16:35.20 When the variance is very high as it often is in species like these social birds.
00:16:42.23 We have a Sage Grouse up on top and a Red-winged Blackbird below.
00:16:47.10 The Sage Grouse is a lekking species. Again, like the elephant seals, a few males get all of the matings.
00:16:54.07 Red-winged Blackbirds are polygamous and so again, one male may have several females in its territory.
00:17:02.07 So the variance in the mating successes is going to be quite high.
00:17:06.17 Some males are going to get a lot of matings. Other males will get very few matings.
00:17:11.02 That will cause the variance to increase.
00:17:14.02 And as you can see from this equation, if the variance increases,
00:17:17.22 the effective population size will decrease.
00:17:21.06 So, again, as is usually the case, the effective population size is much less than the census size;
00:17:29.07 much less than the actual number of individuals who breed at least once in the population.
00:17:34.18 Finally, our last example is how the effective population size changes,
00:17:42.05 if the population size is changing over time.
00:17:46.08 What you can see here are four populations sampled at four different times.
00:17:51.15 Each one has a census size; N sub 1, N sub 2, N sub 3 etcetera.
00:17:58.26 You can see that the size of this population is changing.
00:18:03.03 In the time intervals 2 and 3, the population is quite small
00:18:07.26 whereas in 1 and 4 its much larger.
00:18:09.27 We can encapsulate all those dynamics and changes of population size into a single number
00:18:19.02 by looking at what's called the harmonic mean of the population sizes that we've sampled.
00:18:26.24 The harmonic mean is illustrated by the equation at the bottom.
00:18:30.08 It turns out that the harmonic mean is dominated by small Ns, by small populations.
00:18:37.14 It's essentially the reciprocal of the average of the reciprocals of the population size.
00:18:43.08 And, the fact that we have bottlenecks at generations 2 and 3
00:18:47.29 tells us that overall, our effective population size is going to be much smaller than the actual sizes.
00:18:56.05 And in this case, our effective population size is going to be dominated
00:19:01.00 by those time intervals when the population size was small.
00:19:04.00 OK, so we've gone through effective population size
00:19:09.20 and it's a really important concept and that's why I spent some time on it.
00:19:13.07 It's a really important concept for measuring genetic diversity.
00:19:17.16 And what we're going to do now is talk about two ways of measuring genetic diversity
00:19:22.08 and remember, genetic diversity is important because
00:19:25.10 it's a basic description of the genetics of a population.
00:19:29.28
00:19:32.05 Our first measure of genetic diversity is going to be called theta.
00:19:35.09 It's the Greek letter Q.
00:19:37.05 Theta is equivalent to 4 times the effective population size times the mutation rate,
00:19:45.10 that's our little mu you can see there.
00:19:47.26 So this number, which is usually a fairly big number,
00:19:51.22 N is often 100 or 1000. Mu might be a fairly small number like 10 to the 16 or 10 to the -7
00:20:00.04 depending on how we're measuring it.
00:20:01.20 Multiply those together, and we get what's called the population mutation rate.
00:20:07.29 Essentially, it's the number of new mutations occurring in a population each generation.
00:20:15.08 And what i've shown you here is a way of measuring this
00:20:19.16 from vary straight-forward genetic data from a population.
00:20:23.10 So, let's say we have sampled 3 alleles, 3 forms of a gene, in a population.
00:20:30.11 What we're going to do is we're going to count the number of differences,
00:20:35.18 the number of single nucleotide differences between those alleles.
00:20:38.21 In this example here, you can see there are 3 differences.
00:20:42.08 So our S is going to equal 3. You can see in that first nucleotide A differs from G in allele 3.
00:20:51.00 Later on we've got allele 1 differing in a C where the others have a T.
00:20:55.25 And then finally, in that second to last site,
00:21:00.04 we've got allele number 3 differing as a C where the others have a T.
00:21:07.00 So, that is S as 3 and
00:21:11.22 we can simply calculate theta by plugging it into this straight-forward equation,
00:21:16.11 where we put a 3 in the numerator
00:21:19.03 and below we're going to calculate what's called a Taylor Series,
00:21:23.20 where we're going to go from i equals 1 all the way up to i equals n minus 1, where n is our sample size.
00:21:33.23 In our case, n is 3.
00:21:36.24 So, we're going to simply divide 3 by this expansion of 1 over 1 plus 1 over 2 plus 1 over 3.
00:21:46.20 And that's a very simple way of measuring this population mutation rate
00:21:52.17 which was developed by Waterson and is also called Waterson's Theta.
00:21:57.19 So it's a very straight forward measure.
00:22:01.23 We can sequence DNA, count the differences,
00:22:04.06 estimate the combination of effective population size and mutation rate.
00:22:09.05 And the important thing to remember is that
00:22:11.26 the mutation rate is always going to be a locus-specific estimate of theta.
00:22:18.17 Some genes will have high mutation rate. Some gene will have a low mutation rate.
00:22:23.29 Most genes under neutrality should all have the same effective population size
00:22:31.11 although we'll see some interesting differences from that later on.
00:22:35.17 But, the N should always be the same from gene to gene.
00:22:39.01 What might differ though is the mutation rate
00:22:41.18 and so it's important to remember that theta is always a gene specific measurement.
00:22:46.16 OK, another way of calculating theta: same parameter but calculated a different way.
00:22:55.27 We're also measuring 4 times N times mu
00:22:59.17 and again we're going to have three alleles in our population.
00:23:03.19 This measure is going to take advantage not of the number of polymorphic sites
00:23:09.12 that you saw in the earlier slide
00:23:11.11 but instead we're going to look at what is called the number of pair-wise differences,
00:23:17.18 the number of differences between each pair of alleles.
00:23:23.11 Now, again, if we sample three alleles, we have three different pairs of alleles.
00:23:29.17 We have 1 versus 2, 1 versus 3, and 2 versus 3.
00:23:34.26 So three different pairs.
00:23:36.14 We can count the number of differences between each of those pairs
00:23:40.23 and those number of differences are indicated by the ks.
00:23:45.19 k sub 1,2; k sub 2,3; k sub 1,3. That's what's in the numerator.
00:23:51.11 In the denominator we can have what is written as three choose two.
00:23:58.02 How many ways are there to choose two objects when we have three?
00:24:03.07 Turns out that that number is simply 3.
00:24:06.02 Right? There's three pairs in our population.
00:24:08.21 So, again, it's a very simple measurement.
00:24:12.23 We're going to take the average of the pair-wise differences among those three pairs.
00:24:18.26 And we're going to...that's basically what it is.
00:24:21.28 The denominator is simply our average. We're going to divide by the total number of pairs.
00:24:27.03 So our overall measurement is simply
00:24:31.07 the average pair-wise difference between all alleles in the population.
00:24:36.23 Even when those alleles are the same, we want to count them as separate copies.
00:24:42.05 So those will enter as 0 for k.
00:24:45.00 And so this is a second way to measure theta and it's very simple.
00:24:51.22 You can sequence DNA from several individuals, look at the genetic differences
00:24:56.16 and it's a very quick window into this combination of mutation rate and effective population size.
00:25:04.07 Now, we won't be getting into it but sufficient to say that
00:25:09.27 there's some very creative ways of comparing pi and theta, our two measures of genetic variation
00:25:18.13 to ask whether or not the population is experiencing natural selection.
00:25:23.26 It turns out pi and theta should be the same under neutrality.
00:25:29.23 You should get the same number.
00:25:31.20 It turns out that if those numbers differ,
00:25:34.09 that's one signal that they could be experiencing some deviation from that ideal model.
00:25:41.22 Maybe that there's natural selection going on.
00:25:46.05 It may be that there's some other deviation like
00:25:48.29 there's population substructure maybe of two populations when you thought you just had one.
00:25:54.18 We won't be going into detail in that measure but it's called Tajima's D.
00:25:58.28 But, it's a very common measure for looking at natural selection
00:26:03.27 when you have genetic variation from natural populations.
00:26:07.25 OK, let's look at some examples of theta or what's also called nucleotide diversity;
00:26:17.25 the amount of genetic variation within a species.
00:26:20.26 You can see on this slide nucleotide diversity measured in a bunch of different mammalian groups.
00:26:26.20 For example, you can see Dasyuromorpha which are a group of marsupials.
00:26:32.00 You can see Carnivora, which are carnivores like cats and dogs.
00:26:37.11 You can see Cetartiodactyla which is the big group to which hoofed mammals and whales belong to.
00:26:46.02 They're actually related to each other.
00:26:48.06 You can see Primates, Chiroptera or bats
00:26:52.01 and what you can see is that nucleotide diversity, on average, differs quite a lot
00:26:57.18 between these different groups.
00:26:59.00 In the rodents, at the far right, you see a vary high genetic variation
00:27:06.21 very high levels of nucleotide diversity.
00:27:09.01 Whereas, in the groups towards the left, the Perissodactyls and the Dasyuromorpha
00:27:15.29 they have very low genetic diversity.
00:27:18.09 And, the question is, why do these groups of mammals differ in their diversity?
00:27:24.00 It turns out there are multiple explanations.
00:27:27.24 Rodents may have a larger N.
00:27:30.29 They may have an larger effective population size
00:27:33.18 and be able to sustain more genetic variation than other groups.
00:27:37.15 It may also be that for some reason the mutation rates in rodents are higher
00:27:44.19 on a per generation basis or on a per year basis than those of other groups.
00:27:51.17 Remember, mutation rate will also influence nucleotide diversity.
00:27:55.18 But, surveys such as these are very important
00:27:59.01 for getting a broad view of genetic variation within different groups.
00:28:04.04 Here's an example comparing the nucleotide diversity in many different mammalian groups to those of birds
00:28:11.28 and what you can see again is that the birds
00:28:14.26 generally have lower genetic diversity within their species than do mammals.
00:28:20.23 This is really really interesting.
00:28:22.24 It could mean that bird populations and their effective population sizes
00:28:28.10 are lower than those in mammals.
00:28:30.12 It may also be that birds have a lower mutation rate
00:28:34.28 overall, across their genome, across all their loci than do mammals.
00:28:40.28 And so, again, teasing out these different sources of differences in nucleotide diversity
00:28:50.18 can be very challenging but it's really an essential part
00:28:53.09 to understanding what this simple statistic might mean.
00:28:56.18 OK, let's talk now a bit about gene trees and another way of looking at genetic variation.
00:29:05.16 One thing we're going to see is that the gene tree is
00:29:10.22 a measure of the relatedness of alleles to each other.
00:29:14.07 We've seen that in some of the earlier slides.
00:29:16.23 And two interesting patterns emerge from decades of studies of gene trees.
00:29:23.08 One is that gene trees...the patterns in the relationships among alleles in a population
00:29:30.13 don't always match the species population histories that they're sampled from.
00:29:37.16 We might think that gene trees should be the same as the relationship of species to each other
00:29:43.12 but in fact they can sometimes be different and we'll see some examples of that.
00:29:47.20 Second, that the divergence times of alleles as they coalesce back into the past,
00:29:55.25 those divergence times can differ from the divergence times of the populations harboring them.
00:30:03.20 It's a very non-intuitive result
00:30:06.10 that was first discovered through empirical analysis of simple molecules like mitochondrial DNA.
00:30:13.02 What we are going to see is how the gene tree patterns
00:30:17.20 sometimes differ from the population histories
00:30:20.28 but how we can use these patterns to understand the nature of population divergence.
00:30:27.04 So what you can see here are two examples of population history.
00:30:33.15 In both examples you have single ancestral populations splitting into three populations.
00:30:40.12 First, splitting into two at time T1
00:30:43.22 and then splitting one of the lineages into another two populations at time T2.
00:30:51.13 What you can see in the image on the left is that the bold lines,
00:30:58.22 i.e. the particular ancestry of alleles we're looking at
00:31:03.03 actually have a set of relationships that's different from the population history.
00:31:08.16 You can see that alleles sampled from species 2 and species 3 down there at the bottom
00:31:15.17 those are actually more closely related to each other in the gene tree
00:31:21.14 than are the two populations that branched off
00:31:25.13 more recently in the scenario which are species 1 and 2.
00:31:30.15 You can see that there's a conflict between the genetic history and the actual population history.
00:31:36.04 Now, on the gene tree on the right,
00:31:38.25 you can see that, in fact, the gene tree history matches the species history.
00:31:45.04 Both the gene tree and the species tree have populations 1 and 2 being most closely related
00:31:52.29 and then population 3 branching off earlier.
00:31:56.05 Now, this was a big surprise to geneticists when they discovered this phenomenon.
00:32:01.24 It's a phenomenon called incomplete lineage sorting.
00:32:06.17 Lineages being the genetic lineages percolating through populations.
00:32:10.22 It turns out that we don't need to invoke any special phenomena
00:32:16.20 like hybridization or gene flow to explain these discordant patterns in the gene trees.
00:32:24.14 We can actually simply invoke a succession of very rapid speciation events
00:32:32.18 such that the genetic variation in the ancestral population doesn't quite catch up,
00:32:38.23 doesn't come to fixation, in population before the next speciation event has occurred
00:32:44.26 and it's a perfectly normal process.
00:32:49.23 It happens all the time in populations especially when they diverge in rapid succession from each other.
00:32:56.09 You can see here these parameters on the far right
00:33:00.27 showing twice the effective population size and then also the time (T) divided by 4Ne.
00:33:12.27 Turns out that if the interval between the speciation events is very, very short,
00:33:19.23 on the order or less than 4N generations, where N again, is our effective population size,
00:33:26.22 we're going to have a good likelihood of seeing this incomplete lineage sorting.
00:33:31.24 The alleles will not have sorted out in time.
00:33:34.19 Another phenomenon that we see is this sort of overshoot of the coalescence time
00:33:40.16 from the time of ancestry of the populations.
00:33:43.20 So you can see, for example, that in both of these scenarios,
00:33:47.04 the ancestry of the alleles (those bold lines)
00:33:51.25 goes way farther back than does the ancestry of the actual populations.
00:33:58.10 So you can see that the common ancestor of all the populations begins at time T1.
00:34:03.10 It's that upper line in both those scenarios,
00:34:06.01 but you can see that the genetic variation, the genetic coalescence, of alleles goes back even farther.
00:34:11.27 Of course it's this coalescence that we see in genetic data when we sample it.
00:34:17.12 And what we want to be aware of is that
00:34:19.19 this is actually a somewhat greater time point of divergence than the actual population history.
00:34:26.19 And the reason that is is because the ancestral population
00:34:30.25 had some genetic variation in it already before it underwent any split.
00:34:36.02 It had some depth to its gene tree. It had some genetic variation.
00:34:41.06 And so some of that is going to persist throughout all these population splits.
00:34:46.28 In fact, we except, on average, to see about 2N extra generations in this gene tree
00:34:55.11 that will overshoot by 2N generations the actual population history.
00:35:01.11 So these are just some interesting patterns that phylogeographers learned about
00:35:06.13 once they started investigating patterns of genetic variation within species.
00:35:10.17 Let's talk about some population processes now.
00:35:14.25 The first I want to point out is what happens when we look at multiple populations
00:35:19.27 and we see a discord between the gene boundaries and the population boundaries.
00:35:27.27 So, again, here's a gene tree.
00:35:30.01 You can see the labels at the top of the tree.
00:35:33.14 Those are the population from which the allele was sampled.
00:35:38.02 You can see population 1, population 2.
00:35:43.10 You can see that this tree is not restricted to the species patterns.
00:35:47.08 There are some alleles from population 2, for example,
00:35:51.14 that are more closely related to alleles from population 1.
00:35:55.17 All of the alleles from population 1 don't cluster together
00:35:59.01 and all the alleles from population 2 don't cluster together.
00:36:03.01 Again, this could have a number of sources.
00:36:07.10 One could be incomplete lineage sorting. That's one possible source of this pattern.
00:36:12.29 Another pattern as we'll see could be gene flow between populations;
00:36:18.20 movement of individuals from population 1 to population 2 and successful breeding in that new site.
00:36:26.24 So, both incomplete lineage sorting and gene flow can cause these discordances
00:36:33.06 between the gene tree and the population tree.
00:36:36.07 We can count the number of discordances in this gene tree
00:36:41.29 by simply moving down the tree and noting when we have a population switch.
00:36:47.16 So you can see where the black dots are.
00:36:49.21 Those are instances where we have a switch from population 1 to 2 or population 2 to 1.
00:36:56.11 And that number is going to useful for understanding the dynamics of these populations.
00:37:03.19 So here is a case where we're going to use s, which you see on the y-axis.
00:37:12.03 This is a different s than what we saw earlier. I hope you're not confused.
00:37:15.29 S is on the y-axis and on the x-axis we see time since divergence of the two populations.
00:37:23.24 That divergence might be very recent, in which case t over N is zero
00:37:29.02 or it may be very big in which t over N is large.
00:37:34.00 But you can see how s, we expect this number of intercalations will be very high
00:37:39.28 at the beginning of the speciation event and it will gradually decline to 1 as the divergence time increases.
00:37:48.11 So this is a way of using this discordance to understand how old two populations are.
00:37:55.21 How long in the past did they split?
00:37:58.02
00:38:00.06 Here's an example from primates where we're going to actually see
00:38:05.09 a discordance between the gene tree and the population tree in real data
00:38:10.07 and in fact data that's very close to our hearts because it involves humans.
00:38:14.19 Here is a gene tree embedded within the actual population history.
00:38:20.20 What you can see is that the gene tree matches the species tree.
00:38:25.00 Humans and chimpanzees are most closely related followed by gorillas and orangutans.
00:38:30.12 You can see that we've looked at the number of generations between the splits indicated by t.
00:38:37.05 So, we have a gene tree that's compatible with the species tree.
00:38:41.13 This is in fact what we see at the majority of genes at the human genome.
00:38:46.18 However, we also see situations where the gene tree doesn't match the species tree.
00:38:53.08 Here's a case where you can see that there's an allele sampled from a chimpanzee
00:38:58.06 which traces back to not an allele from humans but an allele from gorillas.
00:39:05.15 We know that gorillas and chimps are not their own closest relatives.
00:39:10.10 Chimps should be most closely related to humans.
00:39:13.06 None the less, some of their genes trace back to an ancestry with gorillas.
00:39:18.15 This again, is entirely expected.
00:39:22.11 It could be due to simply rapid speciations as we talked about before.
00:39:26.11 However, scientists are increasingly thinking that hybridization in the past
00:39:33.00 could have lead to such patterns.
00:39:35.07 So it's still a point of debate among scientists.
00:39:39.26
00:39:43.11 Again, we're going to count the number of these interspecific coalescence events, our s.
00:39:48.17 This time not to look at the age of populations but to look at
00:39:52.20 how much gene flow, how much movement of individuals is going on between populations.
00:39:57.16 Here we're going to use s as an index of gene flow.
00:40:01.12 Again, you can see s on the y-axis
00:40:04.10 and you can see our measure of gene flow,
00:40:07.08 which is N (effective population size) times m,
00:40:11.29 which in this case is the fraction of the two populations which are exchanging migrants.
00:40:19.11 So m is going to be something like .2 or .4. Again, N is going to something like 100 or 1000.
00:40:28.14 Multiply those together, we get the average number of migrants
00:40:33.10 exchanged between populations per generation.
00:40:37.02 So a number like .5 or 1 or 2.
00:40:41.26 You can see as s increases, our estimate of gene flow (Nm) also increases.
00:40:49.02 Although, it gets challenging to distinguish moderate levels of gene flow
00:40:53.19 where Nm is about 20 from very high levels of gene flow where Nm is say 40.
00:40:59.23 It's very tough to tell the difference between those two with population genetic data.
00:41:04.13 OK, let's look at an example where we are using s to look at gene flow in the next slide.
00:41:11.07 Here we have some work that I did in Australia
00:41:16.19 on a group of birds called Grey Crowned Babblers.
00:41:19.01 You can see I have sampled these populations
00:41:22.03 throughout many regions of Western and Northern Australia.
00:41:25.23 You can see that the colors of each population on the map
00:41:30.09 match the colors on the gene tree.
00:41:33.12 And you can also see, on the gene tree, that not all the colors form individual clusters.
00:41:41.24 The red cluster, for example, falls in two different places.
00:41:46.09 You can see the orange lineages towards the bottom of the tree
00:41:50.24 falling in multiple different places.
00:41:52.23 You can see the light blue lineages not forming a single group.
00:41:57.27 That's lineage H.
00:42:00.20 So, these could all be due to either incomplete lineage sorting or gene flow.
00:42:06.23 In this particular case, gene flow seemed like the best explanation
00:42:11.20 and we can use our s statistic to estimate Nm or the amount of gene flow between these populations.
00:42:19.17 And of course this gene flow will erode the monophyly, the distinctness of each population.
00:42:27.08 Let's talk about another measure of genetic variation and that is Fst.
00:42:34.22 Fst is what's called a fixation index
00:42:37.23 developed by Sewall Wright in the 1930s and 40s.
00:42:42.15 We can use our measures of genetic variation, theta or pi, to estimate Fst.
00:42:50.09 In this case, we've got multiple populations.
00:42:53.23 We can use the theta calculated between different populations (that's our subscript b)
00:43:00.25 or theta calculated within individual populations (that's our subscript w).
00:43:06.07 The ratio of those tells us about how much genetic variation is found among different populations.
00:43:15.05 If there's a lot...if every population is very different, Fst is going to be very high.
00:43:21.10 It's going to be close to 1.
00:43:23.00 By contrast, if every population is similar, is very much the same,
00:43:29.10 all having the same genetic patterns, Fst is going to be very low; close to zero.
00:43:35.18 We can use these measures of diversity to estimate this apportionment of genetic diversity.
00:43:42.24 So, imagine this is the total variation within a population.
00:43:48.22 If we look at individual populations, and find just a small amount of genetic variation
00:43:56.14 with the consequence that most of the genetic variation
00:44:00.20 is actually found between different populations;
00:44:03.24 populations are quite different, Fst is going to be very high, close to 1.
00:44:08.13 By contrast, again if we imagine the total amount of genetic variation within a species,
00:44:16.02 now let's look at individual populations within that species,
00:44:20.24 we might actually find a lot of variation within the individual populations,
00:44:24.25 almost as much as we would find in the entire species.
00:44:28.03 In that case, the amount of variation between different populations is very low.
00:44:33.28 And in this case, Fst would be close to zero, very low.
00:44:38.26 So it's a basic measure of how distinct different populations are.
00:44:43.01 We can use Fst to ask about natural selection.
00:44:49.15 It turns out that if everything is neutral,
00:44:53.06 we expect Fst to be the same for most or all of the genes in the genome.
00:44:59.12 And, we can use so called outliers, Fst outliers, to target individual genes
00:45:07.27 which might be subject to natural selection and might be deviating from neutral processes.
00:45:13.29 So, here's some fictitious diagrams showing
00:45:19.26 how Fst would be expected to change with the average allele frequency.
00:45:25.11 We won't worry about the details here,
00:45:27.06 but this dotted line on the panel to the right indicates the expected, maximum Fst
00:45:34.01 given a particular allele frequency for a particular locus.
00:45:39.05 Each dot in this diagram is a different gene from the same set of populations.
00:45:44.15 What you can see is that above that dotted line we have two red dots
00:45:49.08 which show an abnormally high Fst given their allele frequencies.
00:45:54.27 These are good candidates for alleles that have been driven apart
00:45:59.12 not just by neutral processes but by natural selection.
00:46:02.24 They show conspicuously high differences in frequency between populations.
00:46:07.29 What you can see on the left here is a distribution of Fsts over different loci.
00:46:15.17 Again, we expect Fst to be the same, pretty much, among different genes with a little noise.
00:46:22.04 Much of population genetics is stochastic variation caused genetic drift and the coalescence process.
00:46:29.19 But we expect there to be a mode, a single peak for Fst.
00:46:34.09 However, genes that are very high in their Fst or very low in their Fst
00:46:41.05 are good candidates for loci under different kinds of natural selection.
00:46:47.06 So let's look at an example.
00:46:50.00 Here's an example from a fish called Killifish.
00:46:52.18 It's very common across the Eastern US.
00:46:55.09 You can see these researchers sampled them from several different sites along the Eastern US,
00:47:01.29 in particular at so called super-fun sites,
00:47:04.26 which are highly polluted areas
00:47:06.12 that may have caused natural selection in these fish because of their high levels of pollution.
00:47:13.00 And what you can see is again a distribution of Fsts across many, many loci.
00:47:21.06 Each point on these graphs is a different gene.
00:47:23.16 You can see the allele frequencies on the bottom of the axis.
00:47:27.23 On the y-axis you can see the Fst
00:47:31.12 and our expected maximum Fst is the line going across from left to right.
00:47:39.13 You can see some conspicuous outliers.
00:47:42.26 For example, locus B109 way at the top in that top panel
00:47:47.26 is showing a very high Fst given it's allele frequencies of about 0.5.
00:47:53.20 This tell us that that gene may be undergoing natural selection
00:47:59.00 possibly driven by the pollution in these super-fun sites.
00:48:03.13 Similarly, on the bottom diagram you can also see a number of outliers
00:48:09.00 that may be subject to natural selection.
00:48:11.17
00:48:14.18 Here are some examples from human data.
00:48:17.17 You can see that in this case these researchers looked at 8500 different places in the genome
00:48:24.06 and they calculated Fst among different human populations.
00:48:28.18 You can see that most of these loci show an Fst of about 0.1.
00:48:33.20 That says that human populations are very, very mixed up.
00:48:38.09 There's not a lot of genetic differences between different human populations or so called human races.
00:48:44.25 However, there's a small number of genes, which you can see towards the left of this diagram
00:48:50.12 which show a very high Fst; Fst's approaching 1.
00:48:56.06 And these, again, are our Fst outliers.
00:48:58.26 Again, most of the human genome shows very little differentiation between populations.
00:49:04.18 Only a few loci show high levels of differentiation.
00:49:08.07
00:49:11.11 Another example of natural selection is when we see a rapid approach to what we call reciprocal monophyly.
00:49:19.10 That's the situation we saw earlier
00:49:20.29 when alleles form separate clusters within different populations or species.
00:49:26.02 And what these researchers have done is they've looked at gene trees
00:49:30.16 of different loci along a chromosome in these European Corn Borers, a type of moth.
00:49:37.19 What you can see is that for many of these loci,
00:49:42.25 the red and the blue lineages are from different populations,
00:49:49.10 for most of them they are very mixed up.
00:49:52.07 There's not a good distinction between the red and the blue populations.
00:49:57.02 However, you can see for the gene tree right in the center,
00:50:00.07 there's a strong differentiation between the red and the blue populations.
00:50:05.22 There is what we call reciprocal monophyly.
00:50:08.22 They are showing separate clusters in those two lineages.
00:50:11.29 This is an indication that natural selection may have driven these alleles
00:50:18.04 to rapid fixation in these two populations
00:50:20.22 and caused shifts in allele frequencies to be sure
00:50:25.29 but also shifts in the genealogical pattern.
00:50:28.29 Whereas most loci show a mixed up pattern, perhaps due to incomplete lineage sorting,
00:50:34.15 the alleles in that panel in the center are showing increased reciprocal monophyly, increased fixation.
00:50:43.16 So again, a signature of natural selection.
00:50:47.13 OK, so we've seen different patterns of genetic variation, different patterns in gene trees.
00:50:58.06 And what I want to end with is the ways in which scientists are using
00:51:01.22 climate history to understand patterns of genetic variation within species.
00:51:05.22 What you can see here are projections of the niche, the ecological niche,
00:51:12.18 of a land snail in Northern Australia projected back in time.
00:51:19.10 So what you can see is tis land snail likes moist, wet habitats such as rainforest
00:51:26.22 and you can see in the dark green areas, the habitats where it lives in Northern Australia.
00:51:32.13 You can see that this extent of rain forest is a patchwork in the present day closer to the right.
00:51:40.13 You can see in the center panel that it was quite extensive about 7000 or 8000 years ago.
00:51:49.00 Where as at the last glacial maximum, or LGM, about 120000 years ago,
00:51:56.10 rain forest was quite restricted.
00:51:58.23 On the extreme left, you can see the areas have showed the most stable patches of rain forest
00:52:06.11 over time, across all the other 4 maps.
00:52:09.05 And what was intriguing was that these researchers found
00:52:13.03 that those areas of Northern Australia showed the most stability in the climate and the habitat
00:52:19.15 also showed the most genetic variation.
00:52:23.07 This is a pattern which has now been seen in several different species, both vertebrates and invertebrates,
00:52:29.24 on different continents such as Australia and South America.
00:52:33.21 And it's one of the ways where climate scientists
00:52:37.05 are using genetics and information from climate history
00:52:41.22 to tell what are the influences on genetic variation within species.
00:52:47.00 Why are some populations more diverse genetically than others?
00:52:52.16 Perhaps it's because those populations have been stable over time,
00:52:56.24 and they haven't experienced a lot of extinction and recolonization
00:53:01.15 which would tend to reduce genetic diversity.
00:53:04.24 So I hope what you've seen today is a nice overview of the ways
00:53:11.16 in which scientists can use genetic diversity to study the history or populations.
00:53:17.27 We've seen some basic measures of genetic diversity.
00:53:21.11 Once scientists realized that alleles could be looked at genealogically,
00:53:25.22 and we could count the differences between alleles,
00:53:28.19 we can find different ways of measuring genetic diversity
00:53:32.02 and getting a window into that very important population genetic parameter, the effective population size.
00:53:38.00 We've also seen how we can use patterns in the gene trees,
00:53:42.12 whether alleles are clustering by population or species,
00:53:46.11 or whether they're mixed up between populations and species,
00:53:49.23 to ask how recently populations have diverged or whether they are experiencing gene flow.
00:53:56.05 Finally, we've seen how we can use statistics, such as Fst, to ask about natural selection.
00:54:05.07 Have certain loci been driven to fixation, to high or low frequency,
00:54:11.02 in a particular population because of natural selection?
00:54:14.16 That's what Fst outliers can show us.
00:54:17.21 And then finally, this example from climate studies
00:54:21.11 showing how climate history can be linked with phylogeography
00:54:26.15 to better understand the determinants of genetic diversity in natural populations.
00:54:32.02 So thank you very much
00:54:34.04 and I hope you've enjoyed our quick tour through gene trees and phylogeography.
00:54:39.04

Part 2: Multilocus Phylogeography of Australian Birds

00:00:01.06 Hi. My name is Scott Edwards,
00:00:02.18 and in this second lecture
00:00:06.05 we'll be continuing our tour of phylogeography
00:00:08.17 -- the exciting field of
00:00:10.28 linking genetic variation with geography.
00:00:13.26 And, in particular,
00:00:15.23 I'll be telling you about how I have applied phylogeographic principles
00:00:19.10 to study the biogeography and evolution of Australian birds.
00:00:24.25 Australia is an amazing continent
00:00:27.22 full of amazing animals,
00:00:30.14 and here we're going to use genetic variation
00:00:33.11 to understand the history of those populations and species.
00:00:37.24 So, what we're going to first talk about
00:00:41.09 is the context
00:00:45.13 and the particular biogeographic setting
00:00:47.10 of the problems I'm studying.
00:00:49.17 We'll go to the savannahs of Northern Australia
00:00:52.12 and look at a region called the Carpentarian barrier,
00:00:54.11 which has been very important
00:00:56.28 for producing biodiversity in Australia.
00:01:00.12 We'll then talk a little bit about
00:01:03.09 the kinds of molecular markers I'm using in my studies.
00:01:05.19 There are many ways to measure genetic variation,
00:01:08.11 and we have chosen to use
00:01:12.03 so-called sequence-based markers,
00:01:13.26 as opposed to markers called microsatellites,
00:01:17.12 and we'll talk about some of the advantages and disadvantages
00:01:20.18 of those two markers for looking at phylogeography.
00:01:25.00 Finally, we'll move on to the fun part,
00:01:27.10 which are the case studies.
00:01:29.08 These are applying the principles
00:01:32.00 to specific groups of Australian birds.
00:01:33.20 Each group has a separate twist to it.
00:01:36.22 The first three groups will be
00:01:40.00 focused on this Carpentarian barrier,
00:01:42.09 and the last group, which are the zebra finches,
00:01:46.00 will be a separate but no less interesting biogeographic situation.
00:01:50.17 And we're going to use genetic data
00:01:52.24 to look at several different aspects of population history.
00:01:57.01 We're going to look at population structure and gene flow,
00:02:01.04 the movement of individuals between populations.
00:02:02.24 We're going to look at
00:02:06.22 divergence times
00:02:09.28 and whether different groups of species have diverged at the same times
00:02:13.08 or at different times.
00:02:15.01 And then, finally, we'll look at
00:02:18.05 using many loci that are linked together on the same chromosome,
00:02:21.08 and we'll learn how this sort of linkage
00:02:23.12 can give us extra information
00:02:26.01 about the history of populations.
00:02:27.27 So, why have I decided to work in Australia.
00:02:32.26 It's not just a fun place
00:02:35.10 where you can meet the Crocodile Dundee
00:02:40.26 -- Australia, actually, is extremely important and useful for understanding phylogeography.
00:02:46.23 It turns out that the Australia environment
00:02:49.13 has changed very gradually over the past 15 million years.
00:02:53.19 Unlike the Northern Hemisphere continents,
00:02:56.11 like Europe or North America,
00:02:59.07 Australia never had extensive glaciation.
00:03:02.25 And so we never had a situation
00:03:06.14 where the glaciers had dominated the continent
00:03:09.05 and then receded,
00:03:11.28 which of course has been the main driver of genetic patterns
00:03:17.02 in animals living in Europe and North America.
00:03:20.09 Australia had a relatively stable environment
00:03:23.13 that changed very gradually,
00:03:25.19 and as a result you can see that the ranges
00:03:29.13 of different groups of birds
00:03:31.13 are remarkably similar
00:03:35.25 -- sort of a chockablock pattern, where in two separate groups,
00:03:38.04 their ranges will add up to those of a third.
00:03:40.21 So, here, you can see, for example, in the top row,
00:03:44.18 ranges of brown and black-tailed treecreepers,
00:03:46.20 how, if you put those together,
00:03:49.02 you sort of add to the range of a completely unrelated group,
00:03:52.04 the babblers, the grey-crowned babbler.
00:03:54.20 You also see the black-throated
00:03:57.20 and long-tailed finches in the middle row...
00:04:02.29 those ranges add up to those of a completely unrelated species,
00:04:05.02 the red-backed wren.
00:04:07.08 So, we'll be looking at all of these groups to show
00:04:10.27 how their histories have been similar and,
00:04:13.00 in some cases, different.
00:04:14.29 Now, the Carpentarian barrier is a region of Northern Australia
00:04:20.01 that has had a lot of environmental change,
00:04:21.24 and it's a region where
00:04:24.16 related populations and species meet,
00:04:27.23 and in many cases have diverged into separate species.
00:04:30.19 Why have they diverged here?
00:04:32.26 Well, it's possible that
00:04:37.12 there's been a lot of desertification,
00:04:39.14 a lot of aridification in this region,
00:04:41.18 such that animals can't live there
00:04:44.16 but are separated to the periphery.
00:04:46.10 It could also be that there's been intrusions of the marine realm
00:04:50.20 -- the sea has actually made a finger into the interior of Australia
00:04:54.15 and caused populations to split.
00:04:56.22 Of course, the region in this part of the world
00:05:00.22 has been very complicated.
00:05:02.24 Australia, of course, was connected to New Guinea, to the North,
00:05:05.18 on several different occasions,
00:05:07.10 such that there was a single landmass.
00:05:09.05 Now, New Guinea is of course an island separated from Australia
00:05:13.04 by the Torres Strait.
00:05:14.29 But, nonetheless, this Carpentarian region
00:05:17.21 has been a very important region for evolutionary change
00:05:20.27 in the Australian continent.
00:05:23.28 And that is borne out by these regions...
00:05:26.02 these trees you see at the bottom of the diagram.
00:05:28.14 These are called area cladograms.
00:05:32.21 They are essentially summaries
00:05:34.26 of the relationships of different populations
00:05:37.03 around Australia.
00:05:38.18 So, you can see, for example,
00:05:40.21 in the area cladogram on the left part of the screen,
00:05:43.28 we've got several regions
00:05:46.19 around the periphery of Australia.
00:05:48.14 You've got KP, which stands for the
00:05:51.08 Kimberley Plateau.
00:05:53.18 You've got AL, which stands for Arnhem Land.
00:05:56.00 You've got, on the Eastern part of the continent,
00:05:58.26 you've got EF for Eastern Forests,
00:06:00.29 AP for Atherton Plateau...
00:06:06.24 these... you can see how all of these regions are related to each other
00:06:10.04 in a genealogical or a phylogenetic context.
00:06:13.01 And what these are are essentially predictive tools.
00:06:15.04 They are summaries of the histories of many different animals and plants,
00:06:19.13 and we can use them as tools to predict
00:06:24.14 what the relationships of populations will be,
00:06:27.15 based on species that we haven't yet studied.
00:06:29.25 Most importantly, you can see how the barrier, B,
00:06:34.17 that Carpentarian barrier,
00:06:36.26 is the deepest split on that area cladogram
00:06:40.08 on the left side of the screen.
00:06:41.24 What you can see is that the barrier, B,
00:06:46.20 is an extremely important barrier,
00:06:48.24 the most basal split in the area cladogram on the left.
00:06:53.24 And this means that it's been a very important driver of diversification
00:06:58.28 in Australian birds and other animals and plants.
00:07:03.00 We've spent a lot of time running around Australia
00:07:06.02 -- it's a great place to work,
00:07:08.04 the animals are fantastic --
00:07:10.04 and this slide hopefully illustrates
00:07:13.18 how phylogeography work takes a lot of time and a lot of travel.
00:07:17.12 We've got to sample different populations,
00:07:20.03 we've got to look at many different regions of individuals species,
00:07:24.06 as well as many different species.
00:07:26.12 And you can see how you can
00:07:29.01 easily fill up several decades with this sort of travel.
00:07:31.21 It's a lot of fun
00:07:35.09 and you learn a lot about the people and the places where you're visiting.
00:07:38.09 It turns out the Australians themselves
00:07:42.13 are also an interesting lot,
00:07:44.10 as you can see from this sign here,
00:07:46.22 "Shut The Bloody Gate",
00:07:49.12 they've very direct, they don't like to beat around the bush.
00:07:51.16 Much of our work is done on
00:07:54.15 very large, privately owned stations
00:07:57.17 where they're running cattle or sheep.
00:08:00.04 These stations are often biodiversity reservoirs,
00:08:03.18 they're very important for harboring populations of animals and plants.
00:08:08.27 And so we always make sure we shut the gate
00:08:10.27 so the livestock don't run away.
00:08:15.09 Now, I want to give you an example
00:08:19.04 of an early phylogeographic study I did
00:08:23.04 on a bird called the grey-crowned babbler,
00:08:24.29 very widespread in Australia.
00:08:27.16 What you can see here is that we've sampled several populations,
00:08:33.10 all over the eastern part of the continent
00:08:36.17 as well as the west,
00:08:38.29 and you can see that Carpentarian barrier, there.
00:08:41.01 These birds, it turns out, are divided into subspecies,
00:08:43.21 they're not completely different species,
00:08:45.20 but they're definitely differentiated across that barrier.
00:08:47.19 So, we sampled widely,
00:08:51.05 and just to sort of fast-forward,
00:08:53.08 I can show you some of the results of a survey
00:08:55.20 of a molecule called mitochondrial DNA --
00:08:57.22 we mentioned this in lecture 1 --
00:09:00.03 where we made a comprehensive gene tree
00:09:04.15 for all these different populations.
00:09:06.17 And as we course through this gene tree,
00:09:09.24 you can see how the populations
00:09:12.18 are related to different geographic areas of Australia.
00:09:15.00 This is really the essence of phylogeography.
00:09:17.10 You can see there's a group of lineages
00:09:21.13 up in the north, in Cape York and southern New Guinea.
00:09:24.24 You can see a group in the central part of the east coast.
00:09:28.00 And then you can see a cluster in the southeast.
00:09:32.02 And this is really very marked geographic
00:09:35.27 structuring in this species.
00:09:37.26 Similarly, in the west, you can see a broad group
00:09:40.28 over the top end of the continent in the north,
00:09:43.07 as well as a clade in the northwest.
00:09:46.22 This is great evidence
00:09:50.24 for strong phylogeographic structure.
00:09:52.14 Now, mitochondrial DNA
00:09:56.09 was the workhorse of phylogeography for a long time.
00:09:58.07 It's a very important molecule for phylogeography
00:10:02.00 because it changes very quickly
00:10:04.17 -- it's got a high mutation rate,
00:10:06.16 you can get a lot of variation within a species.
00:10:09.04 On the downside,
00:10:11.26 it's only one molecule.
00:10:14.23 It really only tells one story
00:10:18.15 of what is really a collection of stories,
00:10:20.21 each told by a separate gene.
00:10:22.11 All of the genes in the genome, it turns out,
00:10:26.00 have correlated histories
00:10:28.17 -- they're similar to each other,
00:10:30.19 although not identical,
00:10:32.26 because of the vagaries of population genetics
00:10:36.10 and stochastic forces.
00:10:38.04 If we only look at mitochondrial DNA,
00:10:40.16 we would really only be able to say
00:10:43.27 anything about that particular gene.
00:10:45.24 What we'd like to do is to
00:10:50.08 look at genes in the nucleus as well
00:10:52.10 and get a broader view of what the actual populations of these birds are doing.
00:10:55.15 So, that's the reason
00:10:58.22 why we've moved on to looking at additional genes in the nucleus.
00:11:03.02 Now, some of the models that we'll be using
00:11:06.09 are specifically designed
00:11:09.06 to take in information from many different genes,
00:11:13.00 not just one gene.
00:11:14.13 And what we'll do now is briefly review some of the different models
00:11:17.12 that are typically used in phylogeographic research.
00:11:20.26 So, here's a basic model.
00:11:23.01 It's a model of a population structure.
00:11:26.08 So, we're going to think about
00:11:30.02 different populations connected to each other partially by gene flow,
00:11:32.20 by movement of individuals between the populations.
00:11:35.24 Under this model,
00:11:39.08 we can also estimate the sizes of the populations
00:11:41.17 and the amount of genetic diversity
00:11:43.24 within each population.
00:11:45.18 So, this is a very important model
00:11:49.06 if you're interested
00:11:51.27 in how connected populations are to each other,
00:11:53.22 and it's...
00:11:57.01 the software that we use, called MIGRATE, is very versatile
00:12:01.02 and it can estimate each parameter independently,
00:12:03.15 provided that you have sufficient data,
00:12:06.04 a sufficient number of different loci.
00:12:10.15 We can estimate the size of the populations
00:12:13.06 as well as the amount of gene flow between.
00:12:16.18 Another population model which is commonly used
00:12:20.10 is the so-called isolation-migration model.
00:12:22.27 In this model, we can imagine a single ancestral population,
00:12:25.22 which you see at the bottom of the slide,
00:12:28.08 splitting into two different populations.
00:12:32.02 Each of these three populations, the one ancestral
00:12:33.21 and the two descendant,
00:12:35.24 have a population size, an effective population size,
00:12:41.25 as we learned about in the first lecture.
00:12:43.12 There may be gene flow between these two descendant populations,
00:12:45.19 and in fact, in the recent versions of this model,
00:12:48.18 there are many more than two populations
00:12:50.26 -- you can have any number.
00:12:54.03 You can also imagine that we can estimate
00:12:57.24 the time that this ancestral population
00:13:00.03 split into two populations.
00:13:02.03 Like the migration model we saw earlier,
00:13:03.23 this isolation-migration model
00:13:06.20 uses information from many unlinked loci,
00:13:10.09 many loci that might be on different chromosomes
00:13:14.04 and have slightly different histories
00:13:16.11 that are all correlated with each other.
00:13:18.21 And you can see the parameters
00:13:21.10 that we're using on the top,
00:13:23.06 many of these parameters that we discussed in the first lecture.
00:13:25.17 There's theta, there's divergence times
00:13:29.04 with mutation rates, there's gene flow, etc.
00:13:31.13 Finally, we can talk about a model that is...
00:13:35.10 doesn't involve any gene flow.
00:13:37.23 In fact, it only involves a series of isolation events,
00:13:41.09 a series of divergence events.
00:13:43.07 You might call this the pure isolation model.
00:13:46.07 In this model,
00:13:49.19 it's essentially the history of a phylogeny,
00:13:51.28 a history of populations that have been diverging
00:13:56.06 from a common ancestor.
00:13:58.03 And you can see that these populations
00:14:00.06 are splitting at various times
00:14:02.04 -- those are the taos on the left side of the slide,
00:14:07.01 2 and 3.
00:14:08.14 You can also see the sizes of the ancestral populations
00:14:12.13 -- you can see the theta 1, 2, and 3.
00:14:15.10 All of these are describing the
00:14:18.22 history of the splitting of these lineages,
00:14:20.20 and this sort of model, I think,
00:14:23.25 is very similar in spirit to the previous two models we saw.
00:14:28.09 Again, we're going to use multiple loci
00:14:30.18 that are not linked to each other,
00:14:33.20 that can be treated as quasi-independent,
00:14:36.27 somewhat correlated.
00:14:41.04 Now, we're going to be using a series of loci,
00:14:45.03 a series of genetic markers
00:14:47.05 that you might call sequence-based markers
00:14:50.05 or, in our case, anonymous loci.
00:14:52.02 These are essentially random bits of the genome
00:14:56.03 that may not even encode a protein,
00:14:58.25 they may just be junk DNA in some cases.
00:15:02.15 Why do we like this?
00:15:05.11 Because it turns out that junk tends to change readily.
00:15:09.29 It's doesn't get stuck in one optimum
00:15:13.10 because it actually has something to do,
00:15:15.20 because it actually has a function.
00:15:17.10 We can focus on the variation in this junk DNA
00:15:24.10 because it's relatively abundant,
00:15:26.05 it's under no constraints,
00:15:28.01 and most any mutation that happens
00:15:30.17 will be just as good as what was there before.
00:15:32.19 So, this is very, very handy.
00:15:35.07 Now, we like these sequence-based markers
00:15:40.17 because we can measure genetic parameters very easily with them.
00:15:43.15 We can, for example,
00:15:45.26 compare nuclear DNA
00:15:49.11 directly to mitochondrial DNA.
00:15:51.05 They're both sampled by using DNA sequences.
00:15:54.21 They're essentially the same units,
00:15:56.27 the same type of information.
00:15:59.04 We can talk about
00:16:02.28 the extent to which these genes
00:16:05.25 mutate one way and then back against the ancestral state.
00:16:07.27 That's a bad thing
00:16:11.04 because it erases the history before it.
00:16:13.21 We call that homoplasy.
00:16:15.23 In nuclear DNA,
00:16:18.04 when you look at it in terms of the sequences,
00:16:20.14 that sort of backtracking and homoplasy
00:16:23.26 is minimized.
00:16:25.13 We can build gene trees.
00:16:27.11 As we talked about in lecture 1,
00:16:28.25 gene trees are a very intuitive way to look at population history,
00:16:32.03 and we can compare the diversities
00:16:34.17 easily across different species.
00:16:36.16 It's very easy to do.
00:16:38.16 These are all in contrast to
00:16:41.12 a very popular marker called microsatellites.
00:16:44.04 Microsatellites are essentially repeated regions of the genome
00:16:48.03 that we assay, not by sequencing,
00:16:51.20 but simply by scoring the number of repeats
00:16:55.29 within a marker.
00:16:57.29 It turns out that microsatellites can undergo
00:17:01.08 a lot of back-mutation
00:17:03.15 -- homoplasy is very common.
00:17:05.10 It also turns out that we look at mitochondrial DNA
00:17:09.17 and microsatellites in the nucleus
00:17:13.29 in completely different ways.
00:17:15.07 It's like apples and oranges.
00:17:18.03 It's hard to compare them to each other.
00:17:19.25 Finally, it's by... if we focus on microsats,
00:17:22.25 we tend to only look at the hypervariable parts of the genome.
00:17:26.17 We get, really, an atypical view
00:17:30.16 of what's happening in the genome as a whole,
00:17:32.24 whereas with sequence-based markers,
00:17:35.01 we can take more of an average,
00:17:36.29 which of course is what we want
00:17:39.09 if we want to make statements about the genome as a whole
00:17:42.02 and about the populations as a whole.
00:17:44.10 So, on the bottom you see
00:17:47.07 a challenge for phylogeography.
00:17:50.03 If we define the loci that we're looking at
00:17:54.29 by looking at a small subgroup
00:17:57.02 -- what we call a panel --
00:17:59.17 and then assay those markers in a larger group,
00:18:01.12 which is our actual study population,
00:18:03.02 which you see on the right...
00:18:05.21 it's not a big deal,
00:18:08.06 because new sequencing techniques are allowing us to look at the genome
00:18:11.05 in a very unbiased fashion.
00:18:13.11 So, this so-called ascertainment bias,
00:18:15.09 this bias that we incur by just looking
00:18:18.19 at a subset of previously known markers,
00:18:21.01 is pretty much going to disappear,
00:18:23.09 so it's a very exciting time.
00:18:25.16 Okay, let's move on to the animals.
00:18:29.20 We're going to look at four case studies,
00:18:32.16 each of which has a slightly different story to tell.
00:18:37.00 We're going to first look at a group of birds called grassfinches,
00:18:40.23 which are distributed up in the north of Australia.
00:18:43.00 Then we'll look at a group of birds called wrens --
00:18:46.15 they're not like the wrens in North America
00:18:49.12 -- these Australian wrens are completely unrelated.
00:18:53.17 It turns out they're called wrens because they cock their tails
00:18:56.27 just like the wrens here in North America do.
00:19:00.15 We're then going to look at a group of birds called treecreepers,
00:19:03.17 which also are intriguing
00:19:05.25 because they also have undergone convergence
00:19:08.10 with species in the Northern Hemisphere.
00:19:10.17 Treecreepers are unrelated to
00:19:16.02 the creepers that we have here in the US,
00:19:18.04 or maybe in Europe or Asia,
00:19:20.02 but they nonetheless climb trees,
00:19:22.15 almost like a woodpecker,
00:19:25.02 and they're very, very similar, in many ways,
00:19:27.10 to the creepers here in the Northern Hemisphere.
00:19:29.22 Finally, we'll move on to zebra finches,
00:19:32.26 which have a slightly different history
00:19:36.01 -- they're not separated by the Carpentarian barrier --
00:19:38.13 but can tell us a lot about the history of the arid interior of Australia.
00:19:43.11 And each of these studies was conducted by
00:19:48.29 different, very talented students and postdocs in my lab.
00:19:52.16 So, here are the grassfinches,
00:19:55.06 you can see there are three species.
00:19:57.09 Peophila is the genus name.
00:19:59.09 We've got the long-tailed finch,
00:20:01.15 you can see there in the northwest,
00:20:03.09 as well as the black-throated finch in the east.
00:20:06.26 That long-tailed lineage is split into two separate lineages,
00:20:09.24 so we have three in total.
00:20:12.05 Now, what we did with this study
00:20:14.20 was to collect about 30 different markers
00:20:17.16 from the nuclear genome,
00:20:19.29 and we wanted to look at the pattern of gene trees.
00:20:23.22 Now, you might remember, from the first lecture
00:20:27.06 we talked about a phenomenon called
00:20:31.01 incomplete lineage sorting.
00:20:33.06 That's a situation where the alleles
00:20:35.19 are not separating out by species fast enough,
00:20:38.01 before the next speciation event arises.
00:20:40.09 And the result is gene trees that look mixed up...
00:20:43.23 that look mixed up
00:20:47.15 relative to the actual species history.
00:20:51.19 We predict, in this system,
00:20:54.07 that the eastern species, the black-throated finch,
00:20:56.24 Peophila cincta,
00:20:58.17 will be the most...
00:21:01.09 the first to branch off, with the two western ones secondary.
00:21:03.11 And let's see what the gene trees say about that.
00:21:08.17 So, here you can see 30 different gene trees,
00:21:10.23 and on their tips they're got 3 letters:
00:21:13.21 a, h, and c.
00:21:15.10 The a stand for Poephila acuticauda,
00:21:18.22 which is in the northwest;
00:21:21.04 h is for hecki, also in the northwest;
00:21:24.17 and the c is for cincta, in the east,
00:21:29.07 the black-throated finch.
00:21:30.15 Now, you can see on the top row that those gene trees,
00:21:32.21 there are about 15 or 16...
00:21:35.06 those match that prediction that I told you about in the previous slide.
00:21:38.28 In those gene trees, the eastern lineage,
00:21:41.03 the black-throated finch, cincta,
00:21:43.19 is the first to branch off,
00:21:45.14 with the two northwest ones branching off secondarily.
00:21:48.09 All is well and good so far.
00:21:50.15 But then you look at the next set of trees
00:21:54.05 in the middle and the lower rows,
00:21:56.01 and it turns out those have different patterns.
00:21:58.06 In the middle row, you can see that the cincta allele
00:22:02.12 is actually more closely related
00:22:06.16 to one of the western clades, the acuticauda one, the long-tailed finch.
00:22:08.23 In the bottom, you can see how
00:22:12.24 it's the hecki allele that's closely related to the cincta allele.
00:22:15.04 Very, potentially, confusing.
00:22:18.10 It doesn't seem to match the population history.
00:22:21.27 Of course, as we talked about in the first lecture,
00:22:24.22 we can reconcile this by
00:22:27.18 thinking about the time it takes for alleles
00:22:30.23 to sort out by species,
00:22:32.09 if speciation has been very rapid.
00:22:34.05 And it turns out we can take
00:22:36.24 all of the information from all of these gene trees,
00:22:40.02 and come up with a summary statement
00:22:43.25 that is consistent with all of these stories.
00:22:46.02 And, as you might expect,
00:22:48.04 this scenario will have, as a characteristic,
00:22:52.14 very rapid and recent speciation events,
00:22:54.26 because that's what causes these conflicting gene tree patterns.
00:22:58.02 So, here is an analysis of that data,
00:23:02.06 where we've applied the pure isolation model, okay?
00:23:07.18 What you can see is a single tree of lineages at the top,
00:23:12.10 with the three taxa in there,
00:23:17.06 and you can see we have a number of parameters on this tree.
00:23:20.16 On the left, you can see the time
00:23:25.05 that we've estimated for the divergences.
00:23:27.08 MYA stands for millions of years ago.
00:23:30.13 You can see the effective population size
00:23:34.01 of each lineage in the past.
00:23:35.21 Quite large populations
00:23:39.04 -- 350,000 for that basal ancestor,
00:23:43.01 and 200,000 for the more recent ancestor.
00:23:46.17 Of course, large population size
00:23:50.09 is a situation where you might expect these alleles
00:23:53.06 to segregate slowly.
00:23:54.23 They won't separate rapidly
00:23:56.18 because a large effective population size
00:24:00.05 means that genetic drift is fairly slow
00:24:02.18 -- it's not changing quickly.
00:24:04.26 You can also see the branching order of these populations.
00:24:08.00 Again, not necessarily in conflict
00:24:10.07 with some of those gene trees that had, f
00:24:15.04 or example, the h lineage and the c lineage
00:24:17.09 next to each other.
00:24:18.29 This is entirely expected if speciation has been rapid.
00:24:22.20 And then, finally, we can then actually
00:24:25.21 place confidence limits, not just on the gene trees,
00:24:28.18 but on the entire population history.
00:24:31.26 So, you can see we have a confidence
00:24:35.08 in the species tree of about 0.88,
00:24:37.21 which is pretty high.
00:24:40.18 We'd like to see 0.95 or maybe even 1 for absolute certainty,
00:24:43.25 but 0.88 is a pretty good...
00:24:46.10 we're pretty certain of this result, and that's exciting.
00:24:53.03 Now, a couple characteristics of this study...
00:24:56.18 it turns out that the more genes we look at,
00:24:59.02 the more accurate our estimates are.
00:25:02.24 Each of these four panels shows one of the parameters in that previous slide.
00:25:06.21 And so, for example, panels A and B
00:25:10.10 indicate those divergence times
00:25:12.27 that we saw on the left of the previous slide.
00:25:16.04 Panels C and D indicate the populations sizes,
00:25:21.07 the effective population sizes,
00:25:23.00 of those two ancestors:
00:25:25.12 the ancestor of the whole clade
00:25:27.25 and the ancestor of the two descendant lineages.
00:25:30.25 You can see in all four of these panels...
00:25:33.07 actually, really just three of them, A, B, and D,
00:25:37.02 the uncertainty, which is indicated by those dotted lines,
00:25:39.23 is decreasing as we add more loci, okay?
00:25:45.02 As we're adding more genes,
00:25:47.12 from 2 genes all the way up to 30,
00:25:49.14 our estimates are becoming more certain,
00:25:51.26 and that's good.
00:25:54.01 We want to get more certain estimates.
00:25:55.27 Panel C,
00:26:00.12 now, this is the population size of that more recent divergence.
00:26:02.07 I'll just go back for a second,
00:26:04.17 you can see the theta-ab, there,
00:26:07.06 right in the middle of that tree...
00:26:08.23 that's the parameter we're trying to estimate,
00:26:12.04 the population size of that ancestor.
00:26:14.10 It turns out that particular parameter
00:26:17.19 is a stubborn one
00:26:19.17 -- it doesn't seem to get more certain
00:26:22.04 as we add more genes.
00:26:24.06 And this is just a fact of many population genetics studies;
00:26:27.06 some parameters are difficult to estimate.
00:26:31.14 Okay, so more genes means more accurate results.
00:26:35.28 We also can see an example
00:26:38.08 of what we talked about in the first lecture,
00:26:40.03 wherein the gene coalescence
00:26:42.18 overshoots the species divergence.
00:26:44.16 Again, why is this?
00:26:46.23 Because the ancestral species
00:26:50.05 had a measure of genetic variation in it already,
00:26:51.26 and it had a certain depth to its gene tree,
00:26:54.25 even before there were any splits involved.
00:26:57.28 What you can see is that the black bars
00:27:00.02 show the estimates of the coalescence time,
00:27:03.29 or gene divergence,
00:27:06.04 in those two splits:
00:27:08.11 the most basal split and the split of the two northwest lineages.
00:27:11.01 You can see that those gene splits
00:27:14.04 are quite a bit deeper than the actual splits
00:27:17.07 that we're interested in,
00:27:19.12 namely the population splits.
00:27:21.07 You can see that, for example, on the left,
00:27:27.18 the gene split is as deep as the Plio-Pleistocene boundary,
00:27:29.12 which is about 1.8 million years ago.
00:27:31.15 Of course, in reality,
00:27:33.23 the split that we're really interested in is the population split.
00:27:36.20 That we estimate to have occurred
00:27:39.09 much more recently,
00:27:41.16 on the order of more like half a million years ago.
00:27:43.14 And so, you can see why
00:27:46.12 it's very important to distinguish between the gene trees,
00:27:48.18 sort of the raw data that we collect from sequences,
00:27:51.26 and the actual inferences of the population histories,
00:27:54.25 which is a different... a different estimate.
00:28:00.20 Alright. Let's move now to the red-backed wrens,
00:28:04.17 a very charismatic species.
00:28:07.24 The males are jet black with a shock of red plumage on the back,
00:28:11.11 and the females are much more drab.
00:28:13.26 Now, the taxonomists had told us that,
00:28:17.11 in this particular species,
00:28:21.15 it turns out that populations up in Cape York,
00:28:23.25 in that finger sticking up in the northeast of the continent,
00:28:28.04 those actually should be related
00:28:32.14 to the populations in the north and the west.
00:28:34.17 Now, that's a conflict with that area cladogram
00:28:37.04 that you saw earlier.
00:28:39.02 The area cladogram, which you can see here on the inset,
00:28:43.00 suggests, in fact, that those Cape York populations
00:28:45.12 should be more closely related
00:28:47.29 to the Atherton Plateau and Eastern Forest populations.
00:28:50.22 So, CY is our Cape York...
00:28:53.12 you can see how it's related to New Guinea
00:28:57.01 as well as all the populations on the east, AP and EF.
00:29:00.12 So, we've got a conflict...
00:29:02.26 area cladograms say Cape York is related to one thing;
00:29:06.15 taxonomy and plumage characteristics
00:29:08.25 suggest that Cape York should be related to the northern populations.
00:29:12.17 What is the answer?
00:29:14.12 So, again, we collected
00:29:17.19 a bunch of different genes,
00:29:19.22 data from different gene sequences,
00:29:21.16 and what you can see on this slide
00:29:24.03 is simply a measure of genetic diversity
00:29:27.12 at each of those loci.
00:29:29.14 We collected up... almost 40 different loci,
00:29:32.08 and what you can see on the y axis
00:29:34.15 is the different categories of loci
00:29:37.02 -- we have anonymous loci,
00:29:39.22 we have introns,
00:29:41.19 which are those spaces between the coding regions of genes
00:29:45.11 -- a little molecular biology for you --
00:29:47.06 and on the x axis
00:29:51.15 we have that measure of genetic diversity, theta, or 4Nmu.
00:29:55.01 You can see how it varies quite a lot between different genes.
00:29:58.24 Like we learned in the first lecture,
00:30:01.21 theta is a locus-specific measure
00:30:03.22 -- it depends on the mutation rate.
00:30:06.04 What you can see, though, is that we have
00:30:08.15 a diversity of different markers to
00:30:10.28 estimate the population history of this species.
00:30:12.26 So, what we did was we used a bit of software
00:30:17.22 called Structure, and what Structure is able to do
00:30:21.04 is to cluster individuals into larger groups,
00:30:24.21 and you run the model
00:30:29.07 depending on different numbers of groups.
00:30:30.27 What you can see in the table above
00:30:33.17 is that this variable, K,
00:30:35.18 that's our number of populations
00:30:38.18 in each of these runs.
00:30:40.00 You can see K might be a single population,
00:30:42.10 it might be 2 populations,
00:30:44.06 it might be 3, 4, or 5.
00:30:46.08 It turns out, in this particular study,
00:30:48.06 3 populations were the best explanation
00:30:54.00 for this data.
00:30:55.25 Now... and you can see those 3 populations
00:30:59.04 in the 3 colors in the dots,
00:31:02.16 which are sort of extracted from the map of Australia
00:31:04.28 with the northeast, northwest,
00:31:07.08 and the central east populations.
00:31:10.13 Now, what you can see, in the diagram on the bottom,
00:31:14.26 is that Structure allows us to put every individual
00:31:18.02 into a higher organization.
00:31:19.22 You can see each column is a single individual
00:31:22.01 and that individual has some of its genome
00:31:25.15 from one population
00:31:27.21 and some from another.
00:31:29.05 And those colors match the populations
00:31:31.14 you see on the top of the slide.
00:31:33.16 So, what this tells us is that,
00:31:38.26 in fact, one major population is consistent
00:31:43.05 primarily of the populations from the north and west, okay?
00:31:45.24 That's the yellow dots.
00:31:48.16 The other population is mostly individuals
00:31:51.06 in Cape York, the orange dots,
00:31:53.13 and well as individuals in the central east, which are the purple dots.
00:31:58.22 And this is really exciting,
00:32:01.24 because this suggests that its rejecting the taxonomy conclusion.
00:32:07.14 It turns out that the Cape York population,
00:32:10.01 the orange, appears, actually,
00:32:12.24 more closely related to the eastern populations
00:32:15.08 than to the ones in the northwest.
00:32:18.22 So, we've been able to test that hypothesis, here.
00:32:20.22 We accept the area cladogram prediction
00:32:24.10 and reject the predictions from taxonomy,
00:32:27.14 so this is a classic use of phylogeographic data.
00:32:31.18 Let's talk about another great group of vertebrates,
00:32:35.23 the treecreepers.
00:32:37.24 These are the ones that climb trees like woodpeckers,
00:32:40.09 even though there are actually no woodpeckers in Australia.
00:32:42.20 And it turns out these are also, like the wrens,
00:32:45.02 they're also what we call cooperative breeders.
00:32:47.08 They have an interesting social system:
00:32:49.11 they have helpers at the nest,
00:32:53.14 birds that could breed on their own,
00:32:55.12 but it turns out they wait and help their parents raise the next brood.
00:32:58.19 Why they do this is a bit of a mystery,
00:33:03.04 but... and you can imagine how this might
00:33:07.11 influence their genetic diversity, depending on how they breed together.
00:33:12.04 Anyway, like the babblers, we're sampled them all across the continent.
00:33:15.03 We then went out and looked at genetic data
00:33:20.02 from a bunch of different anonymous loci.
00:33:22.04 Here are some examples of the anonymous loci
00:33:24.18 in the gene trees.
00:33:26.17 What you can see here are...
00:33:28.22 each one is 'AL', for anonymous locus...
00:33:30.25 you can see that there are red lineages,
00:33:32.28 which are the brown treecreeper from the east.
00:33:35.23 You can see there are green lineages;
00:33:38.13 these are the black-tailed treecreepers from the west.
00:33:40.11 You can see that each of the gene trees is different;
00:33:45.13 none of them are identical,
00:33:47.11 they're all showing slightly different patterns.
00:33:50.23 And, nonetheless,
00:33:53.10 one postulate of phylogeography
00:33:55.28 is that if all of these loci are neutral,
00:34:00.12 and they probably are neutral
00:34:02.25 because none of these loci actually encode a protein...
00:34:06.05 if they're all neutral,
00:34:08.12 they should all record the same history.
00:34:11.10 So, even though the gene trees are different,
00:34:14.12 we can imagine we're trying to construct
00:34:16.29 an overarching story
00:34:20.07 that accommodates all of the variation here.
00:34:22.07 Okay.
00:34:24.04 We can talk about a model of gene flow and isolation,
00:34:31.09 the MIGRATION model that we mentioned earlier.
00:34:33.11 We talked in the first lecture about
00:34:37.03 sort of a parsimony approach,
00:34:38.29 where we counted these interspecific coalescent events down the tree
00:34:41.22 -- those were those black dots
00:34:44.17 on those gene trees that we saw earlier.
00:34:46.08 This approach was useful
00:34:48.18 when it was first brought in in the late 1980s.
00:34:50.21 It turns out it makes a number of assumptions
00:34:53.08 that we'd really rather do without.
00:34:56.08 For example, it assumes that our gene tree
00:34:59.15 is known with complete certainty, which is rarely the case.
00:35:01.17 It turns out that it's somewhat cumbersome
00:35:05.08 to test different hypotheses under this framework.
00:35:07.23 And the many different sources of uncertainty
00:35:12.11 in population genetics
00:35:14.24 are not adequately accounted for.
00:35:16.26 So, instead of that parsimony approach,
00:35:19.03 we're going to use an approach
00:35:21.15 called maximum likelihood.
00:35:23.02 This is a more robust statistical approach
00:35:26.22 that's embodied in software such as MIGRATE.
00:35:30.06 And, in this approach,
00:35:33.01 the gene trees are estimates.
00:35:35.14 We're not putting all our money on a specific gene tree;
00:35:38.29 there's going to be some uncertainty around each gene tree.
00:35:42.10 We can also estimate the sizes
00:35:45.27 of each population individually,
00:35:47.18 as well as the amount of gene flow between those populations.
00:35:50.27 In the earlier models,
00:35:53.01 we could really only assume that all the populations
00:35:55.27 were essentially the same size.
00:35:57.25 Finally, the maximum likelihood method
00:36:01.02 is much easier in terms of hypothesis testing.
00:36:04.20 We can rank different hypotheses
00:36:07.02 and ask ourselves,
00:36:09.13 which one's the best explanation for the data?
00:36:13.11 So, we're going to apply MIGRATE to our treecreeper data.
00:36:16.29 You can see the ranges, here,
00:36:19.04 of the brown treecreeper in the east
00:36:22.06 and the black-tailed treecreeper in the west.
00:36:24.16 You can see...
00:36:27.15 we've sampled all over the continent.
00:36:29.13 Now, let's first look at the population sizes.
00:36:32.18 Here are our estimates of theta, or genetic diversity,
00:36:37.17 within each of these populations.
00:36:39.19 And what you can see is that some populations,
00:36:42.05 such as the one in Weipa, in the north,
00:36:44.23 have very little genetic diversity.
00:36:46.21 This suggests that their effective population size
00:36:51.05 is relatively small.
00:36:52.25 Probably not because mu is small,
00:36:55.05 because we imagine that all these populations
00:36:58.00 probably have the same mutation rate for these genes.
00:37:00.21 What's probably small is N...
00:37:03.28 remember our equation 4Nmu.
00:37:05.26 Probably, these birds are isolated,
00:37:07.26 they may not be getting any new diversity
00:37:10.21 from other populations,
00:37:12.08 and as a result their effective population size
00:37:15.16 is fairly small.
00:37:17.26 What you can also see, on the west,
00:37:20.23 is a great diversity of population sizes.
00:37:23.22 The Fitzroy Crossing population size is quite big,
00:37:27.12 perhaps because it's getting genetic input
00:37:31.10 both from the north,
00:37:35.00 from the Douglas Hot Springs,
00:37:36.27 as well as from the south...
00:37:39.20 the southwest, from the Fitzroy Crossing region...
00:37:42.05 I'm sorry, from the Newman region, here.
00:37:45.07 Regardless, the beauty of this model
00:37:49.26 is that we can treat each populations independently;
00:37:52.05 each one has a separate amount of genetic diversity.
00:37:55.23 Let's look at gene flow now,
00:37:59.06 those are the arrows connecting these populations.
00:38:01.16 You can see that, in the east,
00:38:03.29 in the brown treecreeper lineage,
00:38:06.10 we've got a fairly constant flow of gene flow
00:38:10.26 across all of these populations, okay?
00:38:13.03 You can see they're well connected by gene flow.
00:38:16.09 By contrast, Weipa, in the north,
00:38:19.05 is somewhat isolated.
00:38:21.01 That's what this thin arrow means.
00:38:22.28 It's not experiencing as many migrants
00:38:25.27 as are the other populations.
00:38:27.18 This might be leading to its low genetic diversity.
00:38:32.26 In the north and the west,
00:38:35.08 you can see that this northern swath really...
00:38:38.18 all of these populations are fairly well connected
00:38:42.08 to each other, okay?
00:38:44.01 They're not significantly isolated from each other.
00:38:46.09 By contrast, the Newman population
00:38:49.22 in the extreme west is very isolated.
00:38:52.27 It's got relatively low gene flow with other populations.
00:38:56.17 So, you can see how we can get
00:38:59.15 a very nuanced picture of this genetic variation
00:39:02.03 using multiple genes
00:39:04.08 and a sophisticated maximum likelihood model.
00:39:07.06 So, now what we're going to do
00:39:11.22 is talk about the population divergence times
00:39:14.17 on the east and west sides of the barrier.
00:39:17.07 Now, the intriguing thing about these results
00:39:19.28 is that the divergence time estimated for,
00:39:24.10 for example, the treecreepers, here,
00:39:26.04 is actually less than...
00:39:29.11 it's about 300,000 years...
00:39:31.11 it's actually less than the divergence time estimated
00:39:34.09 for a species which is considered
00:39:37.02 a single species, namely the red-backed wrens.
00:39:40.00 Okay?
00:39:41.20 We would expect the treecreepers, perhaps,
00:39:43.25 to have an older divergence time.
00:39:45.24 After all, they look quite different, as you saw in those slides.
00:39:49.08 By contrast, the red-backed wrens
00:39:51.16 are essentially a single species,
00:39:53.16 albeit one that has diverged slightly.
00:39:56.18 So, this is one of the surprising results of this particular study.
00:39:59.23 But, what we'd like to do is
00:40:02.29 throw all of this information together to see...
00:40:06.04 ask the question,
00:40:08.10 do we have a single divergence event going on among these three lineages?
00:40:12.16 Or, do we actually have
00:40:14.26 three separate divergence events?
00:40:16.17 This is the kind of hypothesis
00:40:19.01 you can test with phylogeographic data.
00:40:21.27 So, let's look at that question.
00:40:23.20 What you can see here is an analysis
00:40:28.02 done by my colleague, Mike Hickerson,
00:40:30.03 where we've tested three hypotheses against each other.
00:40:34.12 Those hypotheses are illustrated on the bottom of the diagram,
00:40:37.24 and they are the hypothesis that there was a single divergence event,
00:40:43.02 that's one, you see on the left,
00:40:44.24 there are two divergence events,
00:40:47.08 or that there were three separate divergence events
00:40:49.06 for these three groups.
00:40:51.09 Now, we're using a technique known
00:40:53.23 as Bayesian statistics,
00:40:55.18 we're not going to go into details here...
00:40:57.10 suffice it to say that, going into the study,
00:40:59.25 we didn't have any particular preference
00:41:02.07 for one of these three hypotheses.
00:41:05.11 And, for that reason,
00:41:07.16 we're going to say our prior, our prior beliefs,
00:41:10.19 equally rest on each of the three hypotheses.
00:41:13.07 That's what you see in the open bars, here,
00:41:16.17 you can see that the probability
00:41:19.04 is essentially equal for each of those three hypotheses.
00:41:21.25 Now, in Bayesian statistics,
00:41:24.23 the answer, or our estimate,
00:41:27.03 is called the posterior,
00:41:29.15 and that you can see in the dark bars.
00:41:32.26 And the data suggest, in fact, that
00:41:37.13 the hypothesis with the most weight
00:41:39.27 is the one that suggests that there was a single divergence event
00:41:43.14 across all of these lineages.
00:41:45.04 This is intriguing, because,
00:41:48.23 again, some of the species, like the treecreepers,
00:41:51.11 were very different... two different species,
00:41:53.19 whereas the red-backed wrens were a single species.
00:41:56.02 We might have expected different divergence times.
00:42:00.16 Given the data we have collected,
00:42:02.08 our best hypothesis is that there was
00:42:05.15 a single event, perhaps intrusion of the sea
00:42:09.12 or a desert... a desertification event,
00:42:11.29 that split all three populations at the same time.
00:42:14.19 This is a really exciting result
00:42:18.18 because it gives us a broader view of some of the forces
00:42:21.12 that may be affecting Australian birds a
00:42:23.16 nd other animals in Australia.
00:42:27.12 Okay. We're going to move, now, to our last example,
00:42:30.25 of using multiple genes
00:42:34.08 to look at phylogeographic history,
00:42:35.26 and in this example we're going to move
00:42:38.13 away from the Carpentarian barrier
00:42:40.18 to the arid interior of Australia.
00:42:42.11 This is the true Australian outback,
00:42:46.26 and you better bring extra tires with you,
00:42:49.25 because if you strike a...
00:42:52.00 if you get a flat tire out there,
00:42:54.24 it may be hundreds of kilometers before you reach
00:42:58.00 the next gas station.
00:43:00.00 So, we're going to study a species
00:43:02.20 called the zebra finch,
00:43:05.03 a very widespread and abundant species.
00:43:07.10 Also, a very important species for two reasons.
00:43:10.05 One, it turns out that zebra finches
00:43:13.05 are a very important model for studying
00:43:16.24 what we call vocal learning,
00:43:18.29 the ability of animals to learn how to speak.
00:43:23.06 For that reason, zebra finches are actually good models for human speech.
00:43:27.15 We can do experiments on zebra finches,
00:43:29.28 of course, that we can't do on humans,
00:43:33.02 and yet zebra finches learn their vocalizations
00:43:36.06 just like human infants learn how to speak.
00:43:39.18 So, we can learn a lot about how humans speak
00:43:43.15 by understanding the neurobiology
00:43:46.18 and the genetics of species like the zebra finch.
00:43:50.08 It's also important because
00:43:52.20 the genome of the zebra finch,
00:43:54.16 the entire collection of DNA letters
00:43:58.21 in its chromosomes,
00:44:00.07 has been decoded.
00:44:02.10 There are about 1.2 billion letters
00:44:06.28 in this bird's genome
00:44:09.05 and we know nearly every one of them.
00:44:10.28 And this gives us a really unprecedented window
00:44:13.22 into using the genetic variation to study phylogeography.
00:44:18.19 So, what we're going to do here
00:44:21.22 is we're going to look, not just at random genes,
00:44:24.15 but at genes that we know are close together
00:44:26.13 or far apart on the chromosome of this species.
00:44:29.21 This is the type of information
00:44:32.21 that we can't get from most species on the planet,
00:44:35.22 because we don't know where their genes are located
00:44:39.02 on the chromosomes.
00:44:41.28 Thanks to the genome project, we know,
00:44:44.15 for example, that locus 1 and 2 are 2000 base pairs apart
00:44:47.23 -- kb stands for kilobases, or 1000 base pairs.
00:44:51.19 And we know that locus 1 and 3 are,
00:44:56.06 in this case, 10000 base pairs apart.
00:44:59.06 This information, we didn't have for the three case studies we just saw,
00:45:05.06 but, as you'll see, it provides us a lot...
00:45:07.06 with very useful information.
00:45:09.01 You can see, in this study,
00:45:11.06 we have actually gone fairly light
00:45:13.28 on the numbers of individuals we've sampled,
00:45:16.07 just simply 8x4 is 32 in the mainland Australia,
00:45:18.23 plus 12 in the island system of Timor in Indonesia.
00:45:24.09 It turns out, for many kinds of phylogeographic questions,
00:45:28.13 again, it's not so much the number of individuals,
00:45:30.16 but the number of loci that gives you
00:45:34.00 more information.
00:45:35.28 So, let's look at some of the patterns
00:45:38.13 that we've seen in some of these...
00:45:40.13 both the continental populations of zebra finches
00:45:43.11 as well as the island populations in Indonesia.
00:45:48.11 So, let's look at genetic diversity,
00:45:51.01 it's a standard measure of phylogeography.
00:45:53.27 You can see nucleotide diversity on the y axis
00:45:57.17 and you can see, here,
00:45:59.18 for each of about 26 loci on the x axis,
00:46:04.25 how much genetic diversity we have.
00:46:07.17 You can see that, in the Australian populations,
00:46:11.29 which are in the black,
00:46:14.18 quite a lot of genetic diversity
00:46:17.02 -- those bars are really, really high.
00:46:18.25 By contrast, the island populations, in orange,
00:46:23.15 have much lower genetic diversity.
00:46:26.01 In fact, for several of those autosomal loci
00:46:30.02 and those loci called Z loci...
00:46:32.18 these are loci on the so-called Z chromosome,
00:46:35.16 it's like the mammalian X and Y chromosomes,
00:46:38.27 birds have a Z and a W chromosomes...
00:46:42.19 you can see that for many of those loci,
00:46:46.12 there's actually no genetic variation, none, zilch, zero, nada.
00:46:50.09 Those are loci that show no sequence variation,
00:46:53.26 even across all of those individuals.
00:46:57.06 And this is important:
00:47:00.13 this gives us information about the history of those island populations,
00:47:03.08 as you'll see.
00:47:04.28 The Z chromosome is interesting,
00:47:07.18 because it is present in two copies in males, for birds,
00:47:13.27 and only one copy in females.
00:47:16.16 And so this is like what we see in mammals,
00:47:20.26 where the Y chromosome is present
00:47:24.11 in just one copy in males and not present at all in females.
00:47:29.10 It turns out that the effective population size
00:47:32.17 of the Z chromosome...
00:47:34.08 we expect that to be smaller
00:47:37.03 than the effective size of genes
00:47:39.24 on the autosomes. Why?
00:47:42.13 Because there's fewer copies out there in the population.
00:47:44.26 For every 4 copies of autosomes,
00:47:47.21 there are 3 copies of the Z chromosome.
00:47:52.04 So, you can see that would reduce the amount of diversity.
00:47:56.07 And that's partly why the diversity in the Z chromosome
00:47:59.22 is lower than what we typically see in the autosomes.
00:48:03.09 Okay.
00:48:06.00 You can see, here, some measures of differentiation
00:48:09.09 between the mainland and the island populations.
00:48:11.02 You might remember our measure of genetic differentiation,
00:48:14.19 called FST, that's the fixation index.
00:48:17.21 When FST is high, close to 1,
00:48:21.04 it's... lots of differences between populations.
00:48:23.04 When it's low, close to 0,
00:48:25.12 it's very little difference.
00:48:26.27 You can see that all of these genes show slightly lower,
00:48:33.22 sort of moderate levels of FST.
00:48:35.15 You can also see that none of the genes
00:48:39.10 showed what we call reciprocal monophyly,
00:48:42.02 namely that each, the island and the mainland populations,
00:48:49.00 were clustering in different parts of the gene tree.
00:48:50.23 In all of these gene trees, they're all mixed up.
00:48:52.14 And this is, again, a standard part of population history.
00:48:56.25 Although it may seem complicated,
00:48:58.28 it could be the result of a simple divergence event
00:49:01.20 in the recent past.
00:49:05.02 Let's use our model of Structure
00:49:09.26 to ask how we can place different individuals
00:49:12.16 into population clusters.
00:49:14.11 What you can see is a large number of individuals
00:49:17.24 from the mainland,
00:49:20.17 which, essentially, are not genetically differentiated
00:49:23.15 -- they fall into a single large group
00:49:26.08 that we can call the mainland, or continental, group.
00:49:28.18 That group shows a remarkable amount of difference
00:49:31.28 from the island group,
00:49:34.23 the Timor samples here,
00:49:36.19 which are like black and white.
00:49:38.29 They're completely differently genetically,
00:49:42.02 even though they didn't show that reciprocal monophyly,
00:49:45.01 it was sort of a modest amount of differentiation.
00:49:46.28 This Structure method is able to
00:49:50.13 pull out the differences between populations
00:49:52.16 and really allow us to see the forest from the trees.
00:49:55.27 So, quite a lot of distinction
00:49:58.20 between the mainland and the island populations.
00:50:01.10 We can estimate from these data
00:50:04.15 some information on the population history
00:50:06.06 of both the mainland and the island populations.
00:50:10.20 So, on the top you see an estimate...
00:50:13.25 this is, again, a Bayesian analysis,
00:50:16.01 we get a distribution...
00:50:18.06 an estimate of the effective population size of the mainland population.
00:50:22.07 It turns out it's a whopping 1,300,000.
00:50:26.27 That's a big population size.
00:50:28.18 What that means is there's a lot of genetic diversity.
00:50:31.08 As a point of comparison,
00:50:34.06 let's ask what the effective population size of
00:50:39.00 the entire human species is,
00:50:42.19 all 7 billion people on the planet.
00:50:44.12 It turns out we look as if, genetically as if,
00:50:48.29 we're only 10,000 individuals.
00:50:51.07 We're that genetically similar.
00:50:53.01 The effective population size of humans
00:50:55.19 is much, much smaller than the effective size of zebra finches.
00:50:57.29 Who knows what the actual population of zebra finches is.
00:51:03.20 It's probably... it could well be 7 billion,
00:51:05.28 they're small birds and there's a ton of them all over.
00:51:08.02 The important point is that zebra finches
00:51:10.28 harbor much more genetic diversity than do humans,
00:51:13.08 and that's why their effective population size is so big.
00:51:16.19 By contrast, the island population, which you see on the bottom...
00:51:20.23 very little genetic variation.
00:51:23.13 Our estimate of the effective population size
00:51:27.22 is therefore much, much lower
00:51:30.01 than in the mainland continent.
00:51:32.05 We could come up with
00:51:35.10 an overarching scenario
00:51:38.01 of what has happened to the zebra finch
00:51:40.25 using this isolation-migration model that we talked about earlier.
00:51:45.12 You can see how this population model
00:51:48.17 is much simpler and more straightforward
00:51:52.02 than all those somewhat confusing gene trees.
00:51:54.05 And so I like to summarize my results
00:51:57.13 in terms of these population histories.
00:51:59.18 What you can see is the current day...
00:52:03.15 at the current day, at the top,
00:52:05.22 you can see our effective population size of the current species
00:52:07.24 -- 1.3 million in the continental populations,
00:52:10.00 18,000 for the island populations.
00:52:11.21 Using the isolation-migration model,
00:52:15.11 we estimated a divergence of these two populations
00:52:17.19 about 2 million years ago,
00:52:19.12 quite an old divergence.
00:52:21.10 Perhaps more remarkably,
00:52:23.16 we estimate a tiny, tiny effective population size
00:52:27.24 for that island population
00:52:30.00 when it first started, 2 million years ago.
00:52:32.05 Who knows,
00:52:35.04 maybe a small band of zebra finches
00:52:37.26 flew to the Lesser Sunda Islands,
00:52:40.10 made it far enough to colonize and reproduce,
00:52:43.27 but only a small flock.
00:52:45.18 Of course, it may not have been 5,000 individuals,
00:52:48.22 remember, these are effective sizes.
00:52:50.15 It may have been a small number.
00:52:52.16 But it gives us a window into the history of the species.
00:52:57.16 It tells us that there was a small group of individuals
00:53:02.00 that colonized these islands a long time ago,
00:53:04.03 and that, since then,
00:53:06.09 these two populations have been experiencing independent evolution.
00:53:09.19 So much so that they're actually...
00:53:11.19 you can see the difference in their plumage,
00:53:13.18 they actually look and behave differently.
00:53:17.22 So, this is a view of the species
00:53:20.20 that we're able to get by looking
00:53:23.03 at these many different loci.
00:53:25.06 Okay, in this last bit,
00:53:27.18 we're going to use the physical distance between genes
00:53:31.01 to actually provide information about the population history.
00:53:34.02 So, we're going to talk about a phenomenon called
00:53:39.29 linkage disequilibrium,
00:53:41.28 and that's essentially the correlation of patterns
00:53:44.29 between different sites in the genome, okay?
00:53:47.19 So, we can talk about disequilibrium
00:53:50.10 between sites that are 1 base pair apart,
00:53:53.07 very close together,
00:53:54.29 or we can talk about disequilibrium
00:53:57.14 between sites that are very far apart,
00:53:59.09 maybe even on different chromosomes.
00:54:01.25 And, in this case,
00:54:03.28 we're going to color those cells
00:54:06.19 that are showing a lot of similarity in their genetic patterns...
00:54:12.02 we're going to color those cells black.
00:54:14.03 Those are, if we travel down this triangle,
00:54:16.25 those are going to be the cells
00:54:19.05 that link two sites at different parts along the chromosome.
00:54:21.26 By contrast,
00:54:25.13 cells that are white or grey
00:54:27.24 are showing very uncorrelated patterns, very different patterns.
00:54:29.24 Now, let's look at the pattern of linkage disequilibrium
00:54:33.00 in the island and the mainland populations of zebra finch.
00:54:36.26 You can see that, in the mainland birds,
00:54:40.09 we see relatively modest linkage disequilibrium.
00:54:43.22 There's not a whole lot of black cells in that triangle.
00:54:46.12 That means there's not a lot of correlation between these sites.
00:54:50.23 By contrast, if we look at the island population,
00:54:55.06 which you see here on the right...
00:54:57.10 lots and lots of linkage disequilibrium.
00:55:00.05 What is going on?
00:55:02.09 Well, linkage disequilibrium
00:55:05.09 can arise from many different processes,
00:55:07.16 one of which is a small population,
00:55:12.23 or a founder effect,
00:55:15.21 when populations are established from a small number of individuals.
00:55:18.21 That can cause a lot of the genetic similarity to be similar
00:55:21.02 and cause different individuals to look the same,
00:55:23.19 causing LD, as it's called,
00:55:27.20 to increase.
00:55:29.07 You can see that we see that in the island populations,
00:55:31.13 perhaps because of this founder effect
00:55:34.08 that we saw in the previous slide.
00:55:36.09 By contrast, we don't see any appreciable linkage disequilibrium
00:55:39.29 in the mainland population,
00:55:42.02 because that population is so large
00:55:44.05 and it's very easy for the different sites
00:55:46.25 to have different independent histories.
00:55:51.03 So, could this be a case of
00:55:56.01 what we call founder effect speciation?
00:55:58.01 Where a small group of individuals
00:56:00.28 colonized the islands and then
00:56:04.24 evolved into a separate subspecies or species?
00:56:07.22 Well, possibly, because we know that founder effects
00:56:11.24 can cause allele frequencies to differ.
00:56:14.25 As we saw, the mainland and island populations were quite different.
00:56:17.13 It can cause linkage disequilibrium to increase,
00:56:20.22 as we saw,
00:56:22.16 and it can create new gene combinations
00:56:24.27 and, potentially, new material for evolution, on which to act.
00:56:29.12 So, it may be that the founder effect in this species
00:56:35.10 has helped to move the species along
00:56:38.23 to different populations,
00:56:42.08 and the island and the mainland populations.
00:56:43.27 And here are some examples, on the bottom,
00:56:45.18 you can actually see the plumage of these two populations are quite different.
00:56:49.15 So, we've been able to look at many different aspects of phylogeography,
00:56:53.04 all of which are using different, multiple genes,
00:56:58.01 and we've seen how using more genes
00:57:01.29 reduces our uncertainty of our estimates.
00:57:03.22 We've seen how we might have some discordance in our gene trees,
00:57:09.26 but, nonetheless, the population histories
00:57:11.24 can be fairly straightforward.
00:57:13.11 And we've seen how Australian birds
00:57:17.06 are characterized by lots and lots of incomplete lineage sorting,
00:57:23.02 but we've still got evidence
00:57:24.29 for similar divergence times
00:57:27.05 and similar population structures.
00:57:30.10 So, it's...
00:57:33.25 despite the confusion that sometimes looking at individual genes can bring us,
00:57:38.13 summarizing the patterns in terms of the population histories
00:57:40.20 can be a very informative approach.
00:57:42.23 And I would just like to thank
00:57:45.26 the many people who have helped me work on these projects.
00:57:48.22 Most of these people have been to Australia
00:57:51.03 and so they've had the...
00:57:53.20 they haven't been working in the lab all the time,
00:57:56.06 they've actually been there to see Crocodile Dundee.
00:57:58.20 So, thank you very much
00:58:01.00 and I hope you've enjoyed our tour through Australian bird phylogeography.
00:58:05.07 Thank you.

Part 3: Genomics of Host-Parasite Co-Evolution: a Tale of Birds and Bacteria

00:00:00.26 Hi. My name is Scott Edwards,
00:00:02.17 I'm a professor at Harvard University.
00:00:04.26 So, what we're gonna do is use genetic markers
00:00:08.01 to study phylogeography a bit,
00:00:10.26 but we're going to go beyond that
00:00:13.06 and study aspects of the organism
00:00:15.19 that tell us about the ways in which it interacts with its environment.
00:00:18.24 Specifically, we're going to study gene expression
00:00:21.23 and how that interacts with, in our case,
00:00:25.05 a bacterial pathogen,
00:00:28.17 a disease that's being inflicted on this bird
00:00:31.27 that you see here, the house finch.
00:00:34.05 This is an area of molecular ecology
00:00:36.29 known as host-pathogen interactions,
00:00:38.28 and these sorts of interactions
00:00:41.10 are very exciting
00:00:43.20 because they offer the possibility of
00:00:46.27 looking at co-evolution,
00:00:49.07 the ways in which a pathogen, for example,
00:00:52.01 might cause the host to evolve,
00:00:54.24 which in turn might cause the pathogen to evolve yet again.
00:00:59.08 So, there's a really exciting bunch of questions
00:01:03.26 that one can ask when one has
00:01:08.25 a system involving a pathogen and a host.
00:01:11.23 So, to introduce the two main characters in our story, here,
00:01:17.13 house finches are a very common species of bird
00:01:21.09 in North America,
00:01:23.14 you can find them from sea to shining sea,
00:01:25.10 from all the way in the west coast,
00:01:27.24 where they're native,
00:01:29.17 to the east coast, where they've been introduced in the 1940s.
00:01:33.19 We're also going to talk about a bacterial pathogen,
00:01:36.10 or disease-causing organism,
00:01:39.11 called a mycoplasma.
00:01:41.16 Mycoplasmas are very small single-celled bacteria,
00:01:45.19 and they're, in this case,
00:01:50.01 very debilitating for the house finches, as we'll see.
00:01:54.06 Here's a brief synopsis
00:01:58.22 of what's happened between house finches and mycoplasma.
00:02:00.23 In about 1994,
00:02:03.10 scientists realized that house finches in the mid-Atlantic states
00:02:07.28 were getting sick.
00:02:09.11 They were getting huge amounts of swelling around the eyes,
00:02:13.00 they were experiencing
00:02:15.20 a, what we call, conjunctivitis,
00:02:18.11 they had trouble breathing.
00:02:20.07 All of these were eventually traced to a pathogen,
00:02:23.17 which actually originated from poultry,
00:02:26.24 from chickens and turkeys,
00:02:29.21 and which is known as Mycoplasma gallisepticum.
00:02:33.23 This is a very common pathogen of poultry
00:02:37.22 and, in this case,
00:02:40.21 underwent a relatively uncommon jump
00:02:43.29 from its normal host to a new host,
00:02:46.29 the house finch.
00:02:49.15 Now, the interesting thing about this scenario
00:02:52.03 is that house finches
00:02:56.02 appear to have undergone evolution in the short time
00:02:58.19 after they got infected by this disease.
00:03:01.10 Nine years later,
00:03:03.22 biologists could detect that house finches were becoming resistant to the pathogens.
00:03:10.20 They were showing less incidence of death,
00:03:12.14 they were showing evidence of having encountered the diseases
00:03:15.01 but having survived,
00:03:16.27 and they were showing a lot lower incidences
00:03:19.26 of the swelling and the respiratory problems
00:03:22.21 that I mentioned.
00:03:24.10 So, the main questions we're interest in are,
00:03:26.20 how has the host,
00:03:28.27 in our case the house finch,
00:03:30.09 evolved in response to this novel pathogen
00:03:33.09 that it's never seen before?
00:03:34.19 Secondly, how has the parasite evolved?
00:03:37.20 Since it's jumped from a host, the poultry,
00:03:42.10 to a new host, the house finch,
00:03:44.08 which is phylogenetically quite distant...
00:03:47.06 these house finches and poultry
00:03:49.13 haven't seen a common ancestor
00:03:52.02 since about 90 million years ago,
00:03:53.29 so it's not a trivial thing to jump
00:03:56.29 from a chicken to a house finch.
00:04:00.15 Here's a recent history of house finches.
00:04:02.20 As I said, they are found in the western US, naturally.
00:04:06.26 They were introduced to the eastern US
00:04:10.03 when the laws changed,
00:04:14.00 such that pet traders were...
00:04:16.21 it was illegal to hold fancy birds as caged birds,
00:04:20.15 so like good biologists they simply opened their cages
00:04:23.23 and let the birds free on Long Island.
00:04:26.09 The result was that, over the next 60 years,
00:04:29.18 these birds spread all the way from
00:04:33.02 New York City clear across to the Rocky Mountains,
00:04:36.00 where they're now abutting their former ancestors, okay?
00:04:43.23 In addition, house finches are found in Hawaii,
00:04:45.23 where they were introduced, presumably,
00:04:47.22 in the 19th century.
00:04:49.11 When Joseph Grinnell visited the Hawaiian islands in the early 1900s,
00:04:53.19 he found house finches on every island,
00:04:55.20 which suggested to him
00:04:58.23 that they had been here for several decades,
00:05:01.17 so his estimate was that they had arrived in the 1870s.
00:05:05.23 Now, I want you to remember these recent time frames,
00:05:08.28 because you'll see how the genetic data
00:05:11.17 gives us a very high-resolution picture
00:05:14.20 of the differentiation of these populations,
00:05:17.05 when we use a lot of markers,
00:05:20.18 even though the time frame has been very short.
00:05:24.24 Okay, here's a movie showing the spread of
00:05:28.27 mycoplasma from its beginning,
00:05:31.03 in the mid-Atlantic states,
00:05:33.15 all the way across the eastern US.
00:05:36.02 It took a mere 5 years
00:05:37.25 to move from a single point of origin
00:05:42.03 all across the eastern US,
00:05:45.00 for this pathogen to spread --
00:05:46.28 extremely fast, but also extremely well documented
00:05:50.23 because the Cornell Lab of Ornithology
00:05:54.27 conducted what's called a citizen science survey,
00:05:56.25 wherein bird watchers
00:06:01.09 and people watching their bird feeders
00:06:03.12 would report in on the internet
00:06:06.03 whether they saw sick house finches.
00:06:07.24 In this way, they were able to track the spread of the disease
00:06:11.01 with unprecedented clarity.
00:06:13.25 It turned out that citizen science
00:06:17.24 is an extremely powerful way for biologists
00:06:21.01 to collect ecological data,
00:06:22.17 and you should look for an opportunity
00:06:25.13 to become involved with a citizen science project.
00:06:29.10 Now, we're going to look at three different time points in this study.
00:06:32.21 The first is what we call
00:06:35.27 the pre-epizootic stage,
00:06:38.23 in the 1980s, before the pathogen arose.
00:06:40.16 You can see that, across the entire US and Hawaii,
00:06:44.11 everything is blue,
00:06:46.07 meaning there's no evidence of disease.
00:06:48.05 Now, fast-forward to 2001,
00:06:53.28 and we can see...
00:06:56.12 we're in a situation where half of the country...
00:06:59.25 half of the birds in the eastern US have been hit by this pathogen,
00:07:04.07 whereas the western US is free.
00:07:06.24 Now, in the 1980s,
00:07:09.06 that's a useful time to look at patterns of genetics,
00:07:12.25 because now the birds have been unaffected by the disease.
00:07:15.25 By contrast, in 2001,
00:07:18.25 we've got the eastern birds,
00:07:20.12 which have been experiencing this disease,
00:07:22.26 now, for about 5 or 6 years,
00:07:26.04 whereas the western birds are still naive.
00:07:28.04 Finally, we can look at a third time period,
00:07:31.00 in the mid... in about 2007,
00:07:34.22 wherein the disease has now jumped the Rocky Mountains,
00:07:38.26 come down south through California,
00:07:40.17 and is essentially free
00:07:44.11 only from a few states in the southwest.
00:07:46.05 Here's a situation where we have Alabama populations,
00:07:49.17 which now have been with the disease
00:07:54.05 for 7 years, at least,
00:07:56.01 whereas Arizona populations
00:07:58.02 still have never seen the disease.
00:08:00.03 As we'll see,
00:08:01.29 these two populations have very different responses
00:08:04.08 when they encounter the pathogen
00:08:08.06 in an experimental setting.
00:08:09.18 Okay.
00:08:12.14 Let's just review these time periods.
00:08:15.26 We've got the pre-epizootic population,
00:08:19.05 based, importantly, on museum specimens
00:08:22.21 that were collected by museum collectors
00:08:26.11 in the 1980s.
00:08:27.21 Museums and their collections
00:08:29.25 have been a very important source of information for us
00:08:32.08 because they essentially provide a window into the past.
00:08:35.10 In our case, into a time
00:08:37.22 when the species had not seen a pathogen.
00:08:41.04 We've got our 2001 phase.
00:08:45.00 In this study, we're going to look at gene expression.
00:08:47.05 We're going to look at how birds from Alabama
00:08:49.24 respond to the pathogen
00:08:51.24 when they've been exposed to it
00:08:54.14 in an aviary setting.
00:08:56.03 Okay?
00:08:57.21 Finally, we're going to look at
00:09:00.29 the 2007 situation
00:09:02.29 and we're going to, again,
00:09:05.05 do a gene expression,
00:09:06.27 this time comparing birds from Alabama and from Arizona.
00:09:10.21 Remember, Arizona, still, at the time of the study,
00:09:14.03 hadn't seen the pathogen,
00:09:15.27 although we actually know now that
00:09:18.17 it's beginning to sweep through Arizona,
00:09:20.06 and we are ready to look at the these birds
00:09:22.23 as the disease is sweeping through
00:09:24.18 -- we're going to catch evolution in action.
00:09:28.22 Okay, so let's look at the first study,
00:09:31.01 a study of genetic variation.
00:09:32.18 We're going to look at some of the same phylogeography principles
00:09:36.00 that we've heard about in the first two lectures.
00:09:39.10 Here, we're going to look at
00:09:41.20 a type of genetic markers called AFLPs.
00:09:43.11 This stands for amplified fragment length polymorphism.
00:09:49.26 It doesn't give us the detail of actual DNA sequences,
00:09:53.13 but it allows us to look at genetic variation
00:09:57.02 in terms of fragments,
00:09:59.12 and what you can see on this gel
00:10:02.14 is the huge diversity of fragments
00:10:04.24 that we're getting from different individuals.
00:10:07.14 Each column on this gel is a different individual,
00:10:09.13 you can see the different populations that we've sampled.
00:10:13.21 We've even sampled two related species
00:10:16.11 to the house finch,
00:10:18.10 the purple finch and the Cassin's finch.
00:10:19.26 In this study, we looked at
00:10:23.11 over 160 individuals from 16 populations.
00:10:26.15 Using the AFLP technique,
00:10:28.06 we were able to look at 363 different markers.
00:10:32.04 This is like having 363 different genes to look at,
00:10:36.14 it's a very useful approach
00:10:38.26 for getting large numbers of markers.
00:10:40.17 It turns out that over half of these,
00:10:44.17 over 60% of these markers,
00:10:46.03 were variable between populations,
00:10:48.13 which gives us a lot of information,
00:10:50.29 and you can see that...
00:10:52.29 you can see these variations directly on the gel
00:10:55.08 if you look at the pattern of banding.
00:10:57.14 Now, remember we talked about FST
00:10:59.17 in the very first lecture.
00:11:01.03 Here you see a pie chart
00:11:03.09 showing a type of FST,
00:11:05.05 called AMOVA.
00:11:06.27 What this is saying is that...
00:11:09.20 imagine the whole circle
00:11:12.07 being all of the genetic variation
00:11:14.09 within the entire species.
00:11:17.13 Now imagine we're going to look
00:11:19.23 within single populations
00:11:21.15 and we're just going to average the amount of variation
00:11:23.18 that we find within each of the 16 populations.
00:11:26.21 It turns out that we can find
00:11:30.02 over 70% of the variation
00:11:32.26 in the entire species
00:11:34.14 when we look at just a single population, okay?
00:11:39.14 What that tells us is that
00:11:41.15 house finches are not strongly structured;
00:11:43.15 if you look at one population,
00:11:45.27 you might as well have looked at them all.
00:11:47.15 Okay?
00:11:49.08 So, 70% of the total genetic variation
00:11:51.11 can be found within a single population.
00:11:55.18 Nonetheless, we find
00:11:58.13 a substantial amount of variation
00:12:00.07 among different populations
00:12:01.29 within each subspecies, about 8%.
00:12:05.10 And, when we look at yet different subspecies,
00:12:09.01 so, for example, the subspecies in the US
00:12:11.25 versus the one in Mexico,
00:12:13.21 we find over 20%.
00:12:15.25 So, there's detectable variation
00:12:18.11 at different levels in the species,
00:12:21.01 but within the US populations
00:12:23.03 it's a fairly uniform species, genetically.
00:12:27.27 Let's look at genetic diversity,
00:12:29.25 this is very important to get a snapshot
00:12:32.03 of the species before the disease struck.
00:12:35.13 On the y axis, you can see
00:12:37.12 an estimate of nucleotide diversity,
00:12:38.28 this is our friend theta, or pi,
00:12:41.10 what we've seen in the previous lectures.
00:12:43.03 What you can see is that
00:12:46.12 the amount of diversity in the introduced populations,
00:12:49.16 in the eastern US and Hawaii,
00:12:51.23 is pretty much the same as the amount of diversity
00:12:56.19 in the native situation in California.
00:12:59.29 What this tells us is that the eastern populations
00:13:04.21 did not undergo a strong founder effect,
00:13:07.00 they did not lose a lot of genetic variation.
00:13:11.14 Now, using other markers, other researchers
00:13:13.21 have found a small but detectable
00:13:17.01 decrease in genetic variation.
00:13:18.15 Personally, I think it's still...
00:13:20.18 the jury is still out.
00:13:22.06 We need to do yet further studies
00:13:23.00 to definitively answer the questions,
00:13:26.23 has the eastern population lost genetic variation?
00:13:30.08 This is important because
00:13:33.18 it may be that the disease can spread faster
00:13:35.27 in populations that don't have
00:13:39.24 as much variation in their genes,
00:13:41.20 because, of course, variation is the stuff of adaptation
00:13:44.23 -- that's how populations respond to changing environments.
00:13:51.15 Now, remember how we looked at that time course
00:13:54.04 and how that eastern population
00:13:56.20 was founded only about 70 years ago,
00:13:58.09 and the Hawaiian ones in the late 19th century.
00:14:03.10 Using these 363 markers,
00:14:06.05 we're able, nonetheless,
00:14:07.28 to pick out these birds with almost complete certainty,
00:14:10.26 despite the fact that they've been
00:14:14.05 separated for less than a century.
00:14:15.19 So, for example, the east...
00:14:17.10 the western birds, at the top,
00:14:19.10 cluster quite distinctly from the eastern birds,
00:14:22.09 in yellow,
00:14:23.27 and the Hawaiian birds, in red.
00:14:25.23 And this shows you the power,
00:14:28.14 the extraordinary power,
00:14:30.07 when one uses a lot of genetic markers.
00:14:32.16 If you showed this plot to
00:14:34.28 an evolutionary biologist, they would tell you...
00:14:38.05 oh, you're dealing with three different species...
00:14:40.02 but of course we know that these populations
00:14:42.03 are really all part of the same species,
00:14:44.04 and having been separated
00:14:47.06 only by several decades.
00:14:48.25 Now, we can look at patterns of evolution
00:14:53.24 in the phenotype of these birds
00:14:55.20 in these two plots,
00:14:57.09 and this is work done by my collaborator,
00:15:00.05 Jeff Hill, at Auburn University in Alabama.
00:15:03.03 What you can see is that, on the left,
00:15:05.06 that the epizootic actually
00:15:09.23 differentially affected males as opposed to females.
00:15:12.21 It turns out that the proportion of males
00:15:14.27 in the population of Auburn, Alabama,
00:15:17.19 declined as the epizootic ensued.
00:15:21.02 This is interesting
00:15:23.07 -- why might males be hit harder than females?
00:15:25.20 Maybe males have more
00:15:30.11 testosterone running through their systems,
00:15:32.01 which might compromise their immune systems.
00:15:34.05 There are lots of reasons why males
00:15:39.10 might be hit harder by a disease.
00:15:42.01 On the right side of the panel,
00:15:44.09 you can see the way in which,
00:15:47.01 within the males,
00:15:48.18 redder birds were actually better survivors,
00:15:53.26 in this case, than were dull birds.
00:15:57.17 House finches come in a variety of colors,
00:15:59.15 from bright red to orange to yellow.
00:16:01.20 It turns out that the percentage of bright red birds in the population
00:16:06.09 actually increased after the epizootic.
00:16:09.04 This might mean that red, in this species,
00:16:12.12 is an indicator of health and fitness.
00:16:16.06 Maybe dull birds were differentially affected,
00:16:18.28 negatively affected,
00:16:20.15 by this disease.
00:16:23.02 Now, paradoxically,
00:16:24.28 it turns out that smaller birds
00:16:29.04 tended to survive the epizootic, rather than larger birds.
00:16:31.19 You can see how these indices of body size,
00:16:35.11 we've got weight and tarsus length...
00:16:37.09 all of the surviving populations
00:16:40.09 were actually smaller than the populations before the epizootic.
00:16:44.07 And, of course, females were affected much less, in general,
00:16:49.10 than were the males,
00:16:51.09 those are the bars in brown.
00:16:53.18 But you can see how, looking at different sections of the population
00:16:56.12 gives us insight into how the population
00:16:59.01 is changing as a result of this epizootic.
00:17:03.02 Alright.
00:17:04.27 Let's look at two studies looking at gene expression
00:17:07.21 in the 2001 period and the 2007 period.
00:17:11.13 We're going to use two different approaches,
00:17:13.26 one an older approach called macroarray,
00:17:18.13 in which we look at gene expression
00:17:20.23 using nylon membranes,
00:17:22.12 and then another called microarray,
00:17:24.24 where we look at gene expression
00:17:26.25 using glass slides.
00:17:28.19 Genomic is such a fast-moving field
00:17:31.03 that even these approaches are being superseded
00:17:33.23 by even more high-resolution approaches,
00:17:35.21 which we're currently using.
00:17:37.13 What we've done here is an experiment
00:17:41.10 where we've taken birds from the wild, from Alabama,
00:17:45.13 and infected them with this mycoplasma pathogen,
00:17:48.19 a particular isolate.
00:17:50.12 Then, we've asked,
00:17:53.15 how have these birds responded to this infection
00:17:56.05 by their gene expression?
00:17:58.13 Each of these bars
00:18:01.00 indicates a comparison of the gene expression
00:18:03.23 of infected birds versus uninfected birds.
00:18:08.07 So, for example, you can see that,
00:18:11.04 in this orange slide...
00:18:13.26 this orange bar showing HSP90
00:18:16.24 -- this is heat shock protein 90 --
00:18:19.24 you can see that the expression of that gene
00:18:24.00 was higher in birds that had been infected by the disease
00:18:28.00 as compared with birds that had not been infected.
00:18:33.26 This tells us that HSP90 is a...
00:18:36.01 one of the response genes to this pathogen,
00:18:39.04 and you can actually see many genes of interest
00:18:41.19 in the slide,
00:18:43.02 all of which paint a portrait of the immune response
00:18:46.01 of house finches to mycoplasma
00:18:47.29 at the molecular level.
00:18:51.07 Now, intriguingly, some genes
00:18:53.06 which we thought would be upregulated,
00:18:55.16 that would increase in expression,
00:18:57.16 because they're involved with the immune response,
00:19:00.19 it turns out that they are actually downregulated
00:19:03.10 when birds get infected.
00:19:05.15 So, for example, in blue, here,
00:19:07.11 you can see this MHC Class II-associated invariant chain,
00:19:12.06 or CD74
00:19:15.04 -- that's a gene which is very important for the immune response --
00:19:17.14 it actually went down in expression in infected birds.
00:19:22.11 You can actually see that the MHC itself...
00:19:24.28 MHC stands for major histocompatibility complex,
00:19:28.16 also known as HLA in humans,
00:19:31.20 a very important part of the immune system...
00:19:34.22 MHC genes, themselves,
00:19:37.29 are actually going down in expression when they're infected.
00:19:39.23 This is a real head-scratcher;
00:19:41.09 why would genes important for a response like this
00:19:45.10 go down in expression?
00:19:46.15 Well, it may be that this pathogen
00:19:49.27 is somehow subverting the host immune system.
00:19:55.15 Maybe this pathogen is causing these genes
00:19:57.28 to actually go down in expression.
00:19:59.26 Alternatively, this may be a way
00:20:03.23 for birds to modulate their immune response.
00:20:07.11 Maybe if they were to turn on every gene
00:20:09.28 it would cause something disastrous,
00:20:13.01 like an auto-immune response,
00:20:15.20 and they don't want that to happen.
00:20:17.06 So, we're currently testing these ideas
00:20:19.27 about why these genes
00:20:22.03 are paradoxically going down in expression.
00:20:25.10 Of course, there are a lot of genes
00:20:27.20 that we don't know about in the house finch,
00:20:29.19 novel genes, but, nonetheless,
00:20:31.14 using this macroarray approach,
00:20:33.12 we got a nice overview about how gene expression changes
00:20:37.27 upon infection.
00:20:40.20 We move forward in technology
00:20:44.00 and we're using a newer approach called microarrays.
00:20:47.02 And, in this approach,
00:20:51.08 we're not using nylon membranes
00:20:52.17 but we're using these tiny glass slides
00:20:55.02 where we print thousands of genes,
00:20:56.28 and we can monitor the expression of each of these genes
00:21:00.19 using special probes and RNA
00:21:03.12 that we have retrieved from the spleens of these birds.
00:21:08.11 And here's just a picture
00:21:12.16 showing what these microarrays look like.
00:21:14.07 They're printed by these tiny robots
00:21:18.01 that print these very small points on
00:21:22.26 in thousands of copies on each slide,
00:21:25.08 and they're a very effective way to measure
00:21:28.19 thousands of genes across the genome.
00:21:31.29 So, in this approach, we're using what's called
00:21:34.10 a common reference design.
00:21:36.06 Basically, in this design,
00:21:38.09 we're going to pool the RNA from all our experimental classes
00:21:41.15 into a single pool,
00:21:44.00 and we're going to compare the gene expression
00:21:46.11 of every individual class
00:21:49.07 to that common reference, okay?
00:21:51.25 You can see our individual classes, here.
00:21:54.07 We've got control birds from Alabama
00:21:58.09 which haven't been infected,
00:22:00.12 we've got birds three days after infection from Alabama,
00:22:03.27 14 days after infection...
00:22:05.25 we've got similar groups for the Arizona birds.
00:22:09.02 We're going to label the control and the infected birds
00:22:13.01 with different dyes
00:22:15.14 and we're going to pool all of the classes
00:22:19.08 into a single class
00:22:21.06 so that we can basically compare
00:22:23.20 the relative expression of any particular combination we want.
00:22:27.28 It's a useful design if you have a small amount of material.
00:22:30.21 We can't measure gene expression
00:22:32.24 on individual birds,
00:22:34.25 but we want to do comparisons among all the classes in our study,
00:22:38.05 so it's a very useful approach for us.
00:22:43.02 Now, we confirmed all of the patterns
00:22:46.03 in this microarray analysis
00:22:48.00 with other sorts of classical approaches.
00:22:49.27 For example, we did Northern blots,
00:22:52.17 which are a way of asking about
00:22:56.29 levels of gene expression.
00:22:58.08 This approach has been around for a long, long time
00:23:00.04 and is a very powerful and robust way
00:23:02.13 of measuring gene expression.
00:23:04.07 Here's an example showing HSP90.
00:23:06.17 We're also going to use quantitative PCR.
00:23:09.01 Again, these are simply measures
00:23:11.25 to confirm the results of our microarray study,
00:23:15.07 and of course confirmation and validation
00:23:18.19 are very important parts of
00:23:21.05 any sort of study in genomics and molecular ecology.
00:23:26.04 Now, let's look at some of the patterns
00:23:28.21 that we see in the different groups of birds
00:23:32.10 that we're looking at.
00:23:34.05 On the y axis,
00:23:36.07 you can see a measure of the number of genes
00:23:37.28 or the number of clones that are
00:23:40.01 showing changes in expression
00:23:41.25 -- number of clones being number of dots on those slides --
00:23:44.08 and on the x axis you can see
00:23:47.01 different comparisons.
00:23:48.15 Now, what you can see on the left, here,
00:23:51.00 is that, at baseline,
00:23:54.01 before any birds have been infected,
00:23:57.03 we are actually seeing differences
00:23:59.24 in levels of gene expression
00:24:02.12 between Arizona and Alabama birds.
00:24:04.00 This is really exciting;
00:24:06.00 this means there is geographic variation
00:24:08.11 in gene expression.
00:24:09.29 Just baseline levels, no pathogen involved,
00:24:13.09 we can see that there are lots of genes that are showing differences in gene expression level,
00:24:17.22 so this is a really interesting aspect
00:24:21.04 of gene expression and geographic variation
00:24:23.15 in this species.
00:24:25.07 Now, when we look at day 3 and day 14 within Arizona.
00:24:31.02 What we're doing here is comparing birds
00:24:34.13 that have been unaffected,
00:24:36.20 that are clean of the pathogen,
00:24:38.25 versus birds that have been affected.
00:24:40.23 You can see that in both day 3 and day 14,
00:24:43.26 quite a lot of genes are showing differences
00:24:45.29 between unaffected and infected birds.
00:24:49.19 That's exciting, that means we've got some information to work with.
00:24:53.17 So, we've got a lot of differences
00:24:56.08 between control and infected birds in Arizona.
00:25:00.01 Remember, Arizona birds
00:25:03.17 have never seen the pathogen in nature before
00:25:05.12 when this experiment was done.
00:25:07.12 Now, let's look at Alabama birds,
00:25:09.14 which have had about 12 years of exposure to the pathogen
00:25:13.12 at the time this study was done.
00:25:16.18 Notice how, on day 3,
00:25:18.18 when we compared day 3 birds with control birds,
00:25:21.16 we actually cannot find any genes
00:25:24.17 that show differences in expression
00:25:29.06 between unaffected and infected birds.
00:25:30.25 This is amazing!
00:25:32.24 Somehow, these Alabama birds
00:25:34.19 have figured out how to cope with this pathogen
00:25:37.23 without changing the expression of any genes that we can detect.
00:25:42.28 When we look at day 14,
00:25:44.20 we see a small number of genes,
00:25:46.22 much smaller than what we see in Arizona.
00:25:50.09 So, this is really cool.
00:25:52.08 This says that Arizona and Alabama populations
00:25:57.00 are responding differently to experimental infection
00:25:59.08 by mycoplasma.
00:26:03.01 This is just another way to look at some of those data.
00:26:05.27 You can see, in the green bars,
00:26:07.24 all of these are levels of expression
00:26:10.08 comparing Arizona to Alabama.
00:26:14.07 So, in the top,
00:26:16.13 you can see all of these immune system genes,
00:26:18.12 which are showing higher expression levels.
00:26:20.08 At baseline, this is... before any infections.
00:26:24.23 In the yellow,
00:26:27.23 you can see those same genes
00:26:29.14 between Arizona and Alabama
00:26:31.15 after they've been infected,
00:26:33.26 and what...
00:26:36.10 again, what you can see, like we saw in the earlier study,
00:26:38.20 these genes are being suppressed,
00:26:40.20 being downregulated, in the Arizona birds.
00:26:43.14 This is a very surprising pattern:
00:26:48.25 Arizona birds showing lower expression
00:26:50.25 compared to controls and compared to Alabama birds.
00:26:53.10 So, maybe, again,
00:26:55.27 the Alabama birds are showing a more robust response,
00:26:59.26 they're not being subverted by the pathogen, perhaps.
00:27:05.20 This shows a long list of all the genes
00:27:09.12 that show differential expression in our study.
00:27:12.25 A large number of genes belong to the immune system,
00:27:15.12 to metabolic processes,
00:27:17.28 lots of different genes
00:27:20.16 with lots of different functions.
00:27:21.29 What I want you to pay attention to
00:27:24.24 is this green column, number 4,
00:27:27.04 which shows that, again,
00:27:29.29 many genes in Arizona
00:27:32.19 are showing lower levels of expression
00:27:34.20 than are genes in Alabama,
00:27:36.27 once they're been infected.
00:27:38.19 This is, again,
00:27:40.25 differential response of these two populations
00:27:43.08 with different histories
00:27:45.04 to experimental infection with this pathogen.
00:27:50.16 Yet another way to look at the data,
00:27:52.25 it never hurts to look at it many different ways,
00:27:55.17 you can see, again, Arizona versus Alabama,
00:28:01.02 the immune genes tend to be downregulated in Arizona birds.
00:28:03.24 By contrast, stress genes,
00:28:07.11 which you can see on the right, there, in blue...
00:28:10.21 or, in purple, I should say,
00:28:13.11 these genes are showing higher levels of expression
00:28:15.24 in Arizona versus Alabama birds,
00:28:18.13 exactly what you might expected
00:28:21.16 for a naive population that had never seen this disease before.
00:28:25.02 So, a really intriguing comparison.
00:28:30.12 Now, let's look at the dynamics of change
00:28:34.01 between these three populations.
00:28:35.28 Remember, in the macroarray study,
00:28:38.01 we looked at birds using...
00:28:41.02 back in 2001, we looked at the Alabama birds,
00:28:45.11 and in the more recent study
00:28:47.27 we looked at birds from Alabama and Arizona in 2007.
00:28:52.03 Let's look at the similarities and differences
00:28:54.20 between all three of these populations.
00:28:57.03 What you can see on the y axis, here,
00:28:59.29 this is the percentage of expression differences
00:29:05.14 as compared with the Alabama birds in 2001, right?
00:29:07.24 So, what you see is that Alabama birds in 2007,
00:29:12.19 in this column here,
00:29:14.18 are showing a lot of differences
00:29:16.18 from the same population
00:29:18.17 a mere 6 years earlier, okay?
00:29:20.20 That's really, really interesting.
00:29:22.18 By contrast, the Arizona birds from 2007
00:29:26.03 show quite a bit more similarity
00:29:30.10 with the Alabama birds
00:29:33.04 back when they had less exposure to this pathogen,
00:29:36.09 back in 2001.
00:29:37.28 This is a strong suggestion
00:29:40.05 that the Alabama birds
00:29:42.11 have undergone evolution
00:29:44.04 -- they've changed between 2001 and 2007.
00:29:48.24 And that is our currently hypothesis,
00:29:51.04 that, in fact, Alabama birds
00:29:53.15 have undergone evolutionary change
00:29:56.00 such that their response in 2007
00:29:58.10 is quite different from the response in 2001.
00:30:01.26 By contrast, Alabama birds in 2007,
00:30:06.00 which had never seen the disease,
00:30:07.24 are much more similar to Alabama birds in 2001,
00:30:11.27 which also had had minimal exposure...
00:30:14.05 interesting comparisons among all three of these populations.
00:30:19.13 This just summarizes those results.
00:30:22.03 The signature of microevolution
00:30:24.15 in gene expression
00:30:27.14 appears to have occurred in Alabama birds,
00:30:29.15 as indicated by this dotted line, here.
00:30:31.02 The solid line indicates
00:30:33.26 these points of similarity
00:30:36.03 between the Alabama birds in 2001
00:30:38.14 and the Arizona birds in 2007.
00:30:40.27 And, of course, as we saw,
00:30:43.05 the Arizona and Alabama birds in 2007
00:30:45.24 were quite different from each other in many respects.
00:30:48.04 All of these comparisons
00:30:50.29 suggest that evolutionary change has occurred
00:30:53.17 in the Alabama population,
00:30:55.08 and is one suggestion that
00:30:58.00 disease can be a potent driver of evolutionary change
00:31:00.29 in bird populations.
00:31:03.01 Finally, we can look at the
00:31:07.25 functional consequences of these gene expression changes.
00:31:11.08 And so, what you see here
00:31:13.29 are two panels showing
00:31:17.02 what we call the MG load,
00:31:19.02 this is simply the number of mycoplasma bacterial cells
00:31:22.16 per cell of host, okay?
00:31:26.04 So, higher bars means
00:31:30.04 higher amounts of disease in the particular bird.
00:31:36.05 You can see that the Arizona birds
00:31:39.07 are sustaining higher levels of mycoplasma
00:31:41.14 than are the Alabama birds.
00:31:43.15 This, again, suggests that
00:31:47.00 they're not as good at dealing with this pathogen.
00:31:48.14 What you can also see here,
00:31:50.13 on the right,
00:31:52.03 is that we can develop a simple function of gene expression
00:31:55.06 to show how that can drive
00:31:58.27 the level of mycoplasma down,
00:32:00.27 as a simple function of these
00:32:04.22 patterns of gene expression.
00:32:06.11 And what this tells us is that
00:32:08.19 these expression differences that we're measuring
00:32:11.18 actually have functional consequences for the disease,
00:32:13.21 and this is really exciting.
00:32:15.19 It may mean that
00:32:18.04 we've come closer to figuring out
00:32:20.24 what genes are actually involved with the immune response to this pathogen.
00:32:23.17 Okay.
00:32:26.12 Now, we're going to end the talk
00:32:29.05 by talking about the other side of the coin,
00:32:31.09 namely, the pathogen.
00:32:33.11 The pathogen is this very simple organism,
00:32:38.11 much simpler than the host,
00:32:40.16 but, nonetheless, as we'll see, evolving in very interesting ways.
00:32:43.10 The genome of Mycoplasma gallisepticum
00:32:46.23 was sequenced in 2003
00:32:51.13 and it is less than 1 million base pairs
00:32:54.10 -- Mb stands for megabases.
00:32:57.01 Now, a million may sound like a lot,
00:33:00.11 but in fact it's a tiny genome.
00:33:02.14 There are very... relatively few bacteria that have smaller genomes than that.
00:33:06.01 By comparison, of course,
00:33:08.15 the human genome is 3 billion base pairs,
00:33:11.26 and the house finch genome is about
00:33:14.07 1.4 billion base pairs.
00:33:16.23 So, this pathogen has
00:33:19.19 a much, much simpler genome,
00:33:21.15 and for that reason we can really look at the entirety of its evolution
00:33:26.08 in great detail.
00:33:29.07 This is a picture of the pathogen from the poultry,
00:33:31.25 sequenced in 2003.
00:33:33.17 And this is a picture of the [mycoplasma from house finch] genome,
00:33:39.14 which we've recently sequenced.
00:33:42.02 You can see it looks quite similar.
00:33:44.03 It's also circular, like the poultry pathogen.
00:33:48.16 You can see the variations, the differences between the house finch
00:33:53.19 and the poultry strains...
00:33:55.15 are visualized on the outer blue circle.
00:33:57.09 There are lots and lots of differences,
00:33:59.06 lots and lots of evolutionary changes occurred.
00:34:04.00 On the red circle in the second-in from the outside,
00:34:08.03 you can see all the variations
00:34:10.24 that we find among the house finch
00:34:15.05 strains that we've looked at.
00:34:16.22 So, what we did in this study
00:34:19.10 was to pick 12 house finch strains at different time points
00:34:22.04 during this epizootic.
00:34:23.21 We picked a few from the very early part of the epizootic,
00:34:26.15 in the mid 1990s,
00:34:28.21 we picked several from 2001,
00:34:33.15 and several from 2007.
00:34:36.15 We've compared those with some new strains from poultry
00:34:39.11 that we've also sequenced,
00:34:40.28 and we've been able to look at
00:34:44.22 many different aspects of the evolution
00:34:46.25 of this pathogen
00:34:48.27 by looking at these sequence comparisons.
00:34:50.24 And, in this slide, you can actually see
00:34:54.00 some of the difference we've found.
00:34:55.24 So, for example, these grey bars, here,
00:34:57.27 like labeled 4 and 5,
00:34:59.25 you can see those are deletions
00:35:01.28 that have occurred in the house finch strain.
00:35:04.12 It amounts to about 50,000 base pairs
00:35:07.25 that simply are not present in the house finch mycoplasma strains.
00:35:10.14 We can also look at
00:35:15.04 these so-called insertion sequence elements, IS elements,
00:35:17.07 which are sort of "jumping genes".
00:35:19.03 You can see that there are some novel positions of those IS elements in the house finch strains,
00:35:24.03 relative to the poultry strains.
00:35:26.15 Now, let's look at our old friend, genetic diversity.
00:35:29.23 You can see genetic diversity on the y axis,
00:35:33.12 and what you can see on the x
00:35:36.11 are different samples,
00:35:38.10 both from poultry and from house finches,
00:35:40.04 at different points in time.
00:35:43.22 The first two columns on the left
00:35:46.05 you can see are chicken and turkey
00:35:48.27 sampled from the mid-1990s,
00:35:50.16 about the time this pathogen jumped to the house finches.
00:35:54.03 They have a lot of genetic diversity.
00:35:56.19 You can see that it's between 0.03 and 0.04;
00:35:59.17 that means that about 4 positions in every 100
00:36:03.02 are different,
00:36:06.12 are variable within those groups.
00:36:08.03 On this scale, you can't even see
00:36:10.20 the genetic diversity of the house finch strains,
00:36:12.24 it's so low.
00:36:14.17 We have to blow it up to a new scale,
00:36:18.28 which you see on the right.
00:36:20.08 And what you can see is that, yes,
00:36:22.09 there is some diversity in the house finch strains,
00:36:24.10 but overall it's about 2% of the diversity
00:36:26.16 that we saw in the poultry strains.
00:36:28.14 And so, clear evidence for a founder effect
00:36:31.10 -- as this pathogen jumped from poultry into house finches,
00:36:35.27 the genetic diversity was lost
00:36:39.07 because only a few pathogens,
00:36:42.12 a few individual bacterium
00:36:44.15 made the jump successfully.
00:36:47.06 You can also see that there's some variation over time
00:36:50.24 in the house finch strains.
00:36:52.13 You can see that this 2007 strain
00:36:54.17 has much lower genetic diversity
00:36:57.27 than do the 1994 or the 2001 strains.
00:37:01.13 So, this shows that there's continuing bottlenecks,
00:37:04.23 continuing founder effects
00:37:07.27 within the house finch mycoplasma lineage.
00:37:16.17 Let's look at the gene tree of mycoplasma isolates, okay?
00:37:22.25 And from this, we're going to make some inferences
00:37:25.12 about whether the population
00:37:27.20 is stable or a constant size,
00:37:29.09 or whether it's expanding, growing in size.
00:37:34.27 We expect from population genetic theory
00:37:38.10 to see a gene tree with fairly long internal branches,
00:37:41.28 such as you see, here, on the left,
00:37:44.06 if the population has been stable, okay?
00:37:49.17 It hasn't been growing or shrinking.
00:37:52.18 If a population has been growing,
00:37:54.28 as you see here on the right,
00:37:56.22 we expect to see a gene tree
00:37:58.23 in which a lot of the coalescent events
00:38:01.02 are crammed into a small temporal window,
00:38:04.28 which turns out corresponds
00:38:08.22 to the time of population expansion.
00:38:11.09 This is because, going backwards in time,
00:38:13.23 it's as if it's going through a bottleneck,
00:38:16.11 and we have a big population
00:38:19.01 here in the present day,
00:38:20.19 all of the lineages are coalescing rapidly
00:38:22.23 as the population is contracting
00:38:25.26 going into the past.
00:38:27.22 So, we can actually model the time
00:38:30.12 and the extent of population growth
00:38:34.12 from of gene trees,
00:38:35.28 such as we're going to make from the pathogen.
00:38:39.22 Here's a so-called skyline plot,
00:38:43.01 developed with some software called BEAST
00:38:46.21 -- B is for Bayesian,
00:38:49.14 we're going to talk about Bayesian statistics again --
00:38:50.28 and you can see that we're going backwards in time,
00:38:54.17 we've got years before present on the x axis,
00:38:58.06 and we've got a measure of effective population size on the y axis.
00:39:03.12 You can see that, going backwards in time...
00:39:07.09 well, actually, let's talk about going forwards in time,
00:39:09.29 it's probably more intuitive...
00:39:11.24 we actually can see some evidence
00:39:13.22 for a population expansion
00:39:16.20 about 17 years ago.
00:39:19.02 You can see the effective population size
00:39:21.05 is somewhere between 10^2-10^3
00:39:24.27 17 years ago,
00:39:26.21 and then it shoots up to about 10^3, or 1000,
00:39:30.29 by about 14 years ago.
00:39:33.27 This is a signature in the genetic variation
00:39:36.19 that's showing evidence for a population expansion.
00:39:40.00 You also see a dip
00:39:44.00 in the estimated effective population size.
00:39:46.22 This could be a bottleneck;
00:39:48.23 it could also be a bout of natural selection,
00:39:52.27 which has caused a reduction in genetic diversity,
00:39:56.03 because one variant has come
00:39:59.12 and swept to fixation in the population.
00:40:01.18 So, very detailed dynamics
00:40:04.19 that we can get from this sort of data.
00:40:07.13 Here is a tree of the bacterial lineages.
00:40:11.10 You can see the lineages from 2007,
00:40:16.03 here at the top.
00:40:17.27 You can see lineages from 2001 in the middle,
00:40:21.25 and then the earlier lineages
00:40:23.28 from the mid-90s below.
00:40:26.13 A couple of interesting things to note.
00:40:29.13 First of all, there is a clear clustering by time period.
00:40:32.05 We're seeing sort of a sequential movement
00:40:36.01 from the mid-1990s,
00:40:38.05 the early-2000s,
00:40:40.09 to 2007.
00:40:43.22 We also see some evidence
00:40:47.11 for intermingling of the poultry and house finch strains
00:40:50.08 early on.
00:40:52.02 This brown lineage, TK 2001,
00:40:54.26 is a turkey strain sampled in 2001,
00:40:58.08 and it is closely related to one of the house finch strains
00:41:03.01 sampled from Tennessee in 1996.
00:41:05.03 This might mean that this lineage
00:41:07.09 has actually moved back into poultry
00:41:09.20 from house finches
00:41:11.23 -- maybe we've got some ongoing exchange going on.
00:41:14.23 You can also see that the poultry lineages,
00:41:17.23 down at the bottom,
00:41:19.15 are really the wellspring
00:41:22.05 from which the house finch diversity arose.
00:41:23.24 So, our founder event
00:41:26.18 occurred right at the base of the house finch lineages.
00:41:32.06 Now, here are a couple of other interesting aspects of this tree.
00:41:36.19 First of all, we estimate a coalescence time,
00:41:39.16 or a point of common ancestry,
00:41:42.19 of these mycoplasma lineages,
00:41:44.14 a fairly long time ago.
00:41:46.26 We estimate over 600 years
00:41:51.13 for the common ancestor of the house finch and chicken strains,
00:41:54.03 the most basal...
00:41:56.03 the reference strain which is at the bottom, there,
00:41:59.21 called Rlow.
00:42:01.17 Clearly, this is way older
00:42:04.05 than the pathogen in the house finches,
00:42:06.26 and what this tells us is that the reference strain,
00:42:08.21 Rlow, is not the source
00:42:11.23 of the current house finch
00:42:15.09 mycoplasmas that are circulating today.
00:42:16.19 Some other strain,
00:42:18.23 which is more closely related to the house finch strains,
00:42:22.03 is the source,
00:42:24.06 perhaps this turkey strain that we mentioned earlier.
00:42:27.25 We also estimate an extraordinarily high substitution rate.
00:42:31.20 Substitution rate is like the mutation rate;
00:42:34.29 it's the number of variants
00:42:37.28 that get incorporated into the population per year or per generation.
00:42:41.08 This measure of 10^-5
00:42:44.13 is very, very high, even for a bacterium.
00:42:46.21 And this shows us that this pathogen
00:42:50.08 is evolving extremely rapidly,
00:42:52.11 probably trying to adapt to its new host.
00:42:57.22 Finally, we see a lot of turnover, genetic turnover,
00:43:02.18 in these successive cohorts.
00:43:04.07 We can see that the 2007 lineage
00:43:06.13 is defined by a number of unique genetic variants,
00:43:09.28 indicated by the star, here,
00:43:11.22 that distinguish it from earlier cohorts,
00:43:14.09 in 2001 and 1994.
00:43:18.03 So, again, this might be a signature
00:43:20.08 of ongoing evolution,
00:43:22.17 adaptation to this new host.
00:43:25.23 Now, we're going to end by talking,
00:43:28.08 not about the host immune system,
00:43:30.04 but about the immune system of the bacterium.
00:43:33.00 This is remarkable...
00:43:35.19 before I started this study,
00:43:37.19 I didn't even really know that bacteria had an immune system.
00:43:40.03 It turns out that they have a set of genes called
00:43:44.22 CRISPR loci -- that stands for Clustered Regularly
00:43:47.25 Interspaced Short Palindromic Repeats.
00:43:50.29 These CRISPRs essentially are the bacterial immune system,
00:43:53.09 helping it fight against viruses and phages,
00:43:57.06 things that are even smaller than a bacterium.
00:44:00.23 It turns out that if a bacterium
00:44:03.20 has a healthy, variable set of CRISPR loci,
00:44:06.21 it can shut down the proliferation
00:44:10.15 of phage and viruses in itself.
00:44:12.25 It can essentially turn them off,
00:44:14.17 prevent them replicating,
00:44:17.16 as you see here on the bottom panel.
00:44:22.01 By contrast, if a cell
00:44:24.18 does not have a healthy, diverse set of CRISPR loci,
00:44:27.15 phage and viruses can proliferate at will,
00:44:31.07 and will eventually infect the entire cell
00:44:34.23 and lyse it, namely, kill it.
00:44:39.09 And so that's what you see in this upper panel.
00:44:41.20 And, suffice it to say
00:44:43.25 that it's adaptive and advantageous
00:44:45.21 for cells to have a lot of these CRISPR loci.
00:44:47.13 And I want you to remember this
00:44:49.19 as we look at what's happened to the CRISPR system
00:44:52.19 in these house finch lineages of the mycoplasma.
00:44:58.11 This shows a diversity of the CRISPR loci
00:45:02.05 in the house finch mycoplasma lineages.
00:45:05.13 You can see the 12 house finch strains
00:45:08.24 that we've looked at from 2007 back to 1994.
00:45:12.18 You can also see, for comparison,
00:45:14.22 a CRISPR loci
00:45:17.16 set in the reference strain of
00:45:21.09 the poultry mycoplasma strain.
00:45:23.24 Now, the poultry strain
00:45:27.17 had over 70 different CRISPR loci,
00:45:30.01 a very large and diverse set.
00:45:32.02 By contrast, the house finch strains
00:45:35.03 had no more than 52.
00:45:37.15 Now, on top of that,
00:45:39.16 what you can see is as we move through time...
00:45:42.01 is that house finch lineages in the mid-1990s
00:45:44.21 had a reasonably complete set of CRISPRs.
00:45:47.28 As we move to 2001
00:45:54.04 they are showing deletions and less diversity.
00:45:56.04 When we finally get to 2007,
00:45:58.13 they're showing incredibly low diversity of their CRISPR repeats.
00:46:02.03 In fact, you can see these black ovals,
00:46:04.13 over towards the left of the slide,
00:46:06.10 these are so-called CAS genes.
00:46:10.06 These are CRISPR-associated sequenced.
00:46:13.04 They are essential for the functioning of the CRISPR system.
00:46:19.01 All of the strains from 1994, 1995, and 2001
00:46:22.20 have CAS genes.
00:46:24.28 By contrast, the CAS genes
00:46:27.20 have actually been deleted from the 2007 strains,
00:46:31.07 indicating that their CRISPR
00:46:33.14 has not only been truncated,
00:46:35.13 but it doesn't work at all.
00:46:37.22 So, this is really remarkable.
00:46:40.07 We're seeing this pathogen jettison
00:46:42.24 its own immune system
00:46:44.25 as it's moved into this novel host, the house finch.
00:46:48.17 Why would a pathogen do this?
00:46:51.16 Well, it may be that
00:46:54.18 it's providing a short-term advantage
00:46:57.02 for the CRISPR system
00:46:59.17 and for the pathogen as a whole.
00:47:01.16 It may be that,
00:47:03.26 by getting rid of the CRISPR system,
00:47:05.18 it can actually replicate faster.
00:47:07.26 I would predict, however,
00:47:10.15 that because it's gotten rid of this system,
00:47:12.27 in the long-term,
00:47:14.22 it's going to catch up with it,
00:47:16.13 and that viruses and phage
00:47:19.05 will be able to infect these lineages,
00:47:20.29 more than they would lineages that have an intact CRISPR system.
00:47:25.12 So, we've looked...
00:47:27.29 we've used genomic techniques
00:47:29.28 to look at a variety of questions
00:47:33.14 in this host-pathogen interaction.
00:47:34.18 We've seen evidence for microevolution,
00:47:36.19 very rapid, small-scale evolution of the house finch,
00:47:41.03 in terms of its gene expression,
00:47:42.22 when we compare the Arizona and the Alabama populations
00:47:45.03 at different time points.
00:47:48.19 We've seen how the parasite
00:47:51.09 has jumped from poultry into house finches,
00:47:54.02 and has undergone a series of genetic changes,
00:47:57.15 including getting rid of its own immune system,
00:48:00.04 as a way of adapting to its new host.
00:48:03.19 Again, a very paradoxical result,
00:48:06.08 but one that may characterize
00:48:08.25 other host-pathogen systems as well.
00:48:11.15 Finally, I want to emphasize that,
00:48:14.24 because we've been able to look through time,
00:48:17.02 at different time points in the system,
00:48:20.01 in the first case looking at museum specimens
00:48:23.13 and using the genetic variation of house finches,
00:48:25.24 later on using archived samples
00:48:29.02 of the birds as well as the pathogen,
00:48:30.28 we've been able to look at the details
00:48:33.23 of the genetic dynamics through time.
00:48:35.15 And this, I think, is
00:48:38.20 the real future for looking at
00:48:42.20 these patterns of genetic variation:
00:48:44.26 combining genomic techniques with well-archived, well-curated,
00:48:48.06 well-documented samples
00:48:50.18 is going to be the key to answering some very exciting questions
00:48:53.24 in molecular ecology and evolution.
00:48:56.17 And I'd just like to thank
00:48:59.09 the large number of very talented people
00:49:01.05 who have helped me in these studies
00:49:03.29 -- people working on the house finch gene expression system,
00:49:08.08 folks working on the pathogen evolution system.
00:49:11.05 In short, science today
00:49:13.18 takes large numbers of collaborations,
00:49:15.22 people and scientists with different scientific skills.
00:49:20.00 So, that's what makes science so much fun,
00:49:25.04 is working with people with different skill sets,
00:49:27.09 with different domains of knowledge.
00:49:29.08 So, thank you very much.

Videos in this Talk
  • Part 1: Gene Trees and Phylogeography
    Part 1: Gene Trees and Phylogeography
    Audience:
    • Researcher
  • Part 2: Multilocus Phylogeography of Australian Birds
    Part 2: Multilocus Phylogeography of Australian Birds
    Audience:
    • Researcher
  • Part 3: Genomics of Host-Parasite Co-Evolution: a Tale of Birds and Bacteria
    Part 3: Genomics of Host-Parasite Co-Evolution: a Tale of Birds and Bacteria
    Audience:
    • Researcher
Speaker: Scott Edwards
Total Duration: 2:42:47
Recorded: July 2011
All Talks in Ecology

Talk Overview

In his first lecture, Dr. Edwards explains that studying gene alleles within different populations or species allows the construction of gene trees showing how the groups are related. The gene trees can be used to link genetic variation to geographic distribution of populations; the study of phylogeography.

In Part 2, Edwards expands his discussion of phylogeography with a focus on Australian songbirds. Edwards changes gears in Part 3, and describes studies from his lab investigating the recent evolution of resistance to the parasite Mycoplasma in House Finches and the co-evolution of genetic changes in Mycoplasma.

Speaker Bio

Scott Edwards

Scott Edwards

Scott Edwards pursued his undergraduate studies at Harvard and was a graduate student at the University of California, Berkeley. He began his research on Babbler songbirds in Australia while he was a graduate student. During his post-doc at the University of Florida, Edwards switched the focus of his studies to immunogenetics and population genetics in… Continue Reading

Playlist: Genomics

Related Resources

Avise, J. C., J. Arnold, R. M. Ball, E. Bermingham, T. Lamb, J. E. Neigel, C. A. Reeb et al. 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics 18: 489–522.

Avise, J. C. 2000. Phylogeography: The History and Formation of Species. Cambridge, MA: Harvard University Press. Hewitt, G. 2000. The genetic legacy of the Quaternary ice ages. Nature 405: 907–913.

Hey, J., and C. A. Machado. 2003. The study of structured populations— new hope for a difficult and divided science. Nature Reviews Genetics 4: 535–543.

Knowles, L. L. 2009. Statistical phylogeography. Annual Review of Ecology, Evolution and Systematics 40: 593–612.

Ray, N., and L. Excoffier. 2009. Inferring Past Demography Using Spatially Explicit Population Genetic Models. Human Biology 81:141-157.

Salzano, F. M., N. J. R. Fagundes, N. Ray, M. Beaumont, S. Neuenschwander, S. L. Bonatto, and L. Excoffier. 2007. Statistical evaluation of alternative models of human evolution. Proceedings of the National Academy of Sciences of the United States of America 104:17614-17619.

Balakrishnan, C. N., J. Lee, and S. V. Edwards. 2010. Phylogeography and phylogenetics in the nuclear age, In P. R. Grant and B. R. Grant, ed. In Search of the Causes of Evolution: From Field Observations to Mechanisms. Princeton: Princeton University Press.

Hickerson, M. J., B. C. Carstens, J. Cavender-Bares, K. A. Crandall, C. H. Graham, J. B. Johnson, L. Rissler et al. 2010. Phylogeography’s past, present, and future: 10 years after Avise, 2000. Molecular Phylogenetics and Evolution 54:291-301.

Edwards, S. V., Cameron Devitt, S. & M. Fujita. “Phylogeography”. In press in Encyclopedia of Theoretical Ecology. A. Hastings and L. Gross (eds.). University of California Press, Berkeley.

Reader Interactions

Leave a Reply

Your email address will not be published. Required fields are marked *