Modelling genes: the backwards and forwards of mathematical population genetics - Alison Etheridge

1
00:00:02,470 --> 00:00:15,170
I. Okay.

2
00:00:15,250 --> 00:00:21,610
Thank you very, very much. I'm slightly worried you're here under false pretences because this morning I looked at my abstract,

3
00:00:22,210 --> 00:00:26,200
and the talk I wrote doesn't really fit the abstract, but it does fit the title still.

4
00:00:26,200 --> 00:00:29,890
So I'm hoping you won't feel you're here lured by false advertising.

5
00:00:30,670 --> 00:00:36,100
So I'm going to talk about mathematical models and population genetics and a good place to start actually seem to be population genetics.

6
00:00:36,100 --> 00:00:40,240
What is it? And it's more than the last 20 or 30 years where mathematics has been important.

7
00:00:40,990 --> 00:00:45,040
So as a subject, a scientific subject is rather young. It's only a century or so old.

8
00:00:45,490 --> 00:00:52,960
And we usually trace it back to what we call the modern evolutionary synthesis, which was when the work of Darwin up here and Mendel were reconciled.

9
00:00:53,560 --> 00:00:57,220
So what were their theories? So his natural selection in a nutshell.

10
00:00:57,730 --> 00:01:03,879
So Darwin's theory says that heritable traits that increase Chris reproductive success will become more common in a population.

11
00:01:03,880 --> 00:01:11,560
So if, for example, being taller makes you more likely to have more children and being taller is a trait that you tend to hand on to your offspring,

12
00:01:11,830 --> 00:01:14,350
then being taller will become more common in the population.

13
00:01:14,350 --> 00:01:21,790
The population will gradually become taller, and for it to work for it to make any difference to a population, it requires variability.

14
00:01:21,910 --> 00:01:25,870
It's no good if everyone is the same and offspring must be similar to their parents.

15
00:01:27,490 --> 00:01:31,120
That variability and heritability are given to us by Mendel.

16
00:01:31,450 --> 00:01:37,390
So what does Mendel say? Mendel says that traits are determined by genes, determined in inverted commas.

17
00:01:37,780 --> 00:01:44,169
Because actually it's a bit more subtle than that. But traits are definitely influenced by genes, and genes occur in different types.

18
00:01:44,170 --> 00:01:48,310
That gives us that variability that we need. And offspring inherit genes from their parent.

19
00:01:48,340 --> 00:01:54,730
So you've got heritability. So why did it take 50 years for anyone to notice that these two theories could be brought together?

20
00:01:55,580 --> 00:02:01,389
And perhaps the main reason is that whereas Darwin tended to concentrate on the natural selection,

21
00:02:01,390 --> 00:02:05,530
acting through the accumulation of lots and lots of very small changes to get something really big.

22
00:02:05,920 --> 00:02:12,010
So these are his famous Galapagos finches. And to get from this beak to this beak was not something that happened overnight.

23
00:02:12,010 --> 00:02:15,310
It was an accumulation of lots of very, very small changes.

24
00:02:15,910 --> 00:02:22,209
Mendel, on the other hand, famously looked at peas and he was interested in traits which really were determined by genes.

25
00:02:22,210 --> 00:02:29,260
So whether a pea is green, yellow, wrinkled or round, whether it has a a green pot or yellow pot,

26
00:02:29,260 --> 00:02:33,969
a constricted pot or inflated pot, these are all things determined just by single genes.

27
00:02:33,970 --> 00:02:38,230
That very discrete is very different sort of variability from Darwin and his finches.

28
00:02:38,710 --> 00:02:45,040
But nonetheless the theories were brought together and they were brought together by mathematics and the mathematicians in question.

29
00:02:45,400 --> 00:02:49,120
We usually attribute anyway the modern evolutionary synthesis to three guys.

30
00:02:49,570 --> 00:02:55,240
So we have Fisher over here, Ray Fisher, civil right, and J.B. Haldane.

31
00:02:56,140 --> 00:03:01,380
So Fisher. Very famous British mathematician or statistician, depending on your culture.

32
00:03:03,060 --> 00:03:06,900
Fisher was interested in the data that was being collected by biometrics,

33
00:03:06,900 --> 00:03:12,420
and so they collected data on things like height and weight of parents and their offspring.

34
00:03:13,110 --> 00:03:17,129
And Fisher noticed that this could all be explained by Mendelian genetics as long as

35
00:03:17,130 --> 00:03:21,620
you allowed a particular trait to be determined by lots and lots of Mendelian factors.

36
00:03:21,630 --> 00:03:28,140
So lots of genes influencing the trait, each having a very small influence and a bit of environmental noise.

37
00:03:28,620 --> 00:03:32,820
And in the process, he actually invented much of modern statistics, and in particular the analysis of variance.

38
00:03:33,150 --> 00:03:37,140
This isn't a usual picture of Fisher, but that's what he would have looked like at the time of the evolutionary synthesis.

39
00:03:37,320 --> 00:03:42,650
Usually we show this grey man with a long white beard. Over here on the right, we've got civil right.

40
00:03:43,220 --> 00:03:43,910
So civil right.

41
00:03:43,910 --> 00:03:51,710
Was it an American who was trained in mathematics and then lowered into biology by a woman whose name I wrote down so I knew would never remember it?

42
00:03:52,370 --> 00:03:57,169
Wilhelmina Antman qui, i have you forgive me for not remembering that, but she's interesting.

43
00:03:57,170 --> 00:04:00,200
She was one of the first women to get a Ph.D. from the University of Chicago.

44
00:04:00,680 --> 00:04:04,820
And while Wright was at Cold Spring Harbour, she lured him into biology.

45
00:04:05,360 --> 00:04:11,480
And he developed a lot of what we now call the theory of genetic drift, which is understanding randomness.

46
00:04:11,570 --> 00:04:15,830
This is a long time before probability was a fashionable mathematical subject,

47
00:04:15,830 --> 00:04:23,600
but he was understanding the randomness in a population just because it's finite, it's inherently stochastic, this nature, the nature of reproduction.

48
00:04:24,140 --> 00:04:28,010
And he also developed notions, things like fitness, landscapes, which we still use today.

49
00:04:29,710 --> 00:04:33,100
This man here, JBS Haldane, Oxford trained mathematician.

50
00:04:33,460 --> 00:04:38,350
And you might guess, looking at the photo that he had perhaps the most colourful of the careers of these three gentlemen.

51
00:04:39,550 --> 00:04:43,720
He wrote an excellent children's book, my friend Mr. Leakey, which I thoroughly recommend to you.

52
00:04:44,500 --> 00:04:50,139
He was married to a very interesting journalist who brought very interesting people into his life and into his household.

53
00:04:50,140 --> 00:04:53,950
And so he left Oxford and travelled the world and finally died in India, very sadly.

54
00:04:55,010 --> 00:05:03,720
But. While these three would certainly have agreed that Mendel and Darwin were very compatible theories and indeed they they reinforced one another.

55
00:05:04,200 --> 00:05:07,560
What they certainly did not agree on was the answer to this question.

56
00:05:08,010 --> 00:05:13,889
So what is the relative importance of the different forces of evolution that are acting on my population?

57
00:05:13,890 --> 00:05:19,290
So natural selection in the sense of Darwin population structure, because we don't all live in a big melting pot,

58
00:05:19,290 --> 00:05:26,220
we're all sort of spread around and we live in different spatial locations and in different, different genetic forms.

59
00:05:26,580 --> 00:05:33,069
And genetic drift, this randomness that right. Set up. And I actually deliberately put Fisher and right.

60
00:05:33,070 --> 00:05:36,690
Rather a long way away from each other because they really did not get on.

61
00:05:36,700 --> 00:05:43,300
They had a very long standing feud. Because while right thought that genetic drift was a very important evolutionary force,

62
00:05:43,720 --> 00:05:46,690
Fisher thought that it would be completely dwarfed by natural selection.

63
00:05:46,930 --> 00:05:51,280
And he and Ford wrote a number of rather aggressive papers against rights theory.

64
00:05:52,280 --> 00:05:58,460
So if these incredibly intelligent and innovative thinkers were unable to cast light on this problem,

65
00:05:58,970 --> 00:06:02,780
why do we think we might be able to shed any new light on it now?

66
00:06:03,470 --> 00:06:09,380
And the answer lies in the data. So right you will notice is holding a guinea pig.

67
00:06:09,740 --> 00:06:16,370
This is not because in the 1930s when this photograph was taken, the Americans used guinea pigs as blackboard erasers as it might appear.

68
00:06:17,090 --> 00:06:27,400
It's actually because he bred guinea pigs, so these guys could only view genetic information indirectly by phenotype and right developed the our

69
00:06:27,410 --> 00:06:34,309
understanding of the way that different coat colours are inherited in guinea pigs and also rats,

70
00:06:34,310 --> 00:06:37,840
rabbits and lots of other other mammals of similar descent.

71
00:06:39,070 --> 00:06:43,600
Nowadays we can view DNA sequences directly, and frankly, our data is a lot less cute.

72
00:06:44,110 --> 00:06:47,480
So here is what geneticists do with modern data.

73
00:06:47,550 --> 00:06:49,570
This thanks to Jonathan McKinney for this.

74
00:06:49,810 --> 00:06:56,290
Actually, if you go to the Department of Statistics, you can see this patent on the ball, on the doors, on the second floor.

75
00:06:57,400 --> 00:07:00,580
This is how you can tell the statistical geneticists at work.

76
00:07:01,420 --> 00:07:04,910
And what it corresponds to is data from 40 different human beings.

77
00:07:04,920 --> 00:07:10,930
They're from the Thousand Genomes Project, in fact, and they all come from an area in Nigeria.

78
00:07:11,230 --> 00:07:17,470
And what's been recorded, this is quite a long sequence of DNA, but all this records is the differences between individuals.

79
00:07:18,040 --> 00:07:23,680
And that's what geneticists there record, the differences between individuals or rather between the DNA sequences and individuals.

80
00:07:24,130 --> 00:07:28,960
And from those differences, they infer something about the way that individuals are related to one another.

81
00:07:29,350 --> 00:07:32,659
And we call those relationships genealogical trees.

82
00:07:32,660 --> 00:07:34,870
And we'll see a lot of those in the rest of the talk.

83
00:07:35,710 --> 00:07:41,110
So as mathematicians, if we want to address that key question about the different importance of the different forces of evolution,

84
00:07:41,110 --> 00:07:48,189
what we need are forwards in time models that say how those forces of evolution would change gene frequencies,

85
00:07:48,190 --> 00:07:51,880
how would they change the frequencies of different genetic types to move forwards?

86
00:07:52,300 --> 00:07:57,910
But then we want to compare that to data, or rather to what geneticists infer from their data.

87
00:07:58,240 --> 00:08:03,640
And so we need to be able to say backwards in time, if a population were evolving according to one of our models,

88
00:08:03,970 --> 00:08:10,720
what would those genealogical trees, what would those systems of relatedness look like in individual sample from our population?

89
00:08:12,020 --> 00:08:15,200
Okay. So let's just have a quick think about backwards in time.

90
00:08:16,340 --> 00:08:19,890
So I said genealogical tree and I'm deliberately saying that not family tree.

91
00:08:19,910 --> 00:08:24,700
So let's try and explain why. If I want to plot my family tree, what do I need?

92
00:08:24,710 --> 00:08:29,720
I need my parents and my grandparents and my great grandparents and so on.

93
00:08:30,170 --> 00:08:33,590
And the number of individuals in each generation is growing really, very, very quickly.

94
00:08:33,590 --> 00:08:36,919
And I think that's quite nicely illustrated by Mike Wallace.

95
00:08:36,920 --> 00:08:42,350
Just why you can find this sculpture in the grounds of Morton College. And thanks to David Lowery for sending me the photo.

96
00:08:43,160 --> 00:08:47,899
After just nine generations, which here are meant to represent generations of academics in maudlin,

97
00:08:47,900 --> 00:08:58,460
with a little self-important and maudlin through 512 leaves on this tree, it doesn't take very long to get to a very big number nine generations, 512.

98
00:08:58,940 --> 00:09:01,040
Now, natural populations are finite,

99
00:09:01,520 --> 00:09:08,570
and so you can't indefinitely go on doubling the number of people in your family tree without running out of individuals to put in your family tree.

100
00:09:08,930 --> 00:09:13,729
So some individuals must occur more than once. And let's see a real example of that.

101
00:09:13,730 --> 00:09:21,590
And I bet I've chosen the rather extreme example. But here is the family tree or the pedigree of King Charles, the second of Spain.

102
00:09:22,400 --> 00:09:29,720
And so here's Charles himself and here's his father. And here are his paternal grandparents and then great grandparents and so on.

103
00:09:30,440 --> 00:09:34,639
And then his his mother and here his maternal grandparents.

104
00:09:34,640 --> 00:09:39,080
And then we see that his great grandparents appear to be duplicated already.

105
00:09:39,710 --> 00:09:45,590
So he's a very extreme case because his mother was his father's niece.

106
00:09:46,260 --> 00:09:50,780
Now, this is really quite extreme inbreeding and it goes on as you go back in the tree,

107
00:09:50,780 --> 00:09:54,350
you'll see there are lots of instances of lineages coming together.

108
00:09:54,380 --> 00:10:02,600
This really isn't a tree. And in fact, it is an extreme case because I'm afraid Charles the second was actually very

109
00:10:02,600 --> 00:10:06,230
seriously handicapped by genetic disease and died without leaving offspring.

110
00:10:07,040 --> 00:10:14,540
So let's try and find a family tree, which is a little less politically inspired, because obviously a lot of the marriages here were not random.

111
00:10:14,540 --> 00:10:24,110
They were so that bits of Spain, state property of Spain and I'm going to move to a very apolitical organism that the snail

112
00:10:24,470 --> 00:10:28,250
and one of the reasons I moving to the snail is because drawing these pictures gets very,

113
00:10:28,250 --> 00:10:33,350
very difficult if you separate your population into males and females and snails hermaphrodite.

114
00:10:33,770 --> 00:10:36,409
But let me assure you, the mathematical models are almost identical,

115
00:10:36,410 --> 00:10:40,610
just much harder to draw with the program that I was using in a hotel room in Paris yesterday.

116
00:10:41,300 --> 00:10:44,360
So it's very carefully prepared, etc.

117
00:10:44,870 --> 00:10:50,629
So here what we've done is we've taken five snails and snails are not monogamous.

118
00:10:50,630 --> 00:10:55,040
So we're supposing that in the previous generation each snail just picks two parents at random.

119
00:10:55,040 --> 00:11:00,920
So for example, this one chooses that parent and that parent and any snail can breed with any other snail because they're hermaphrodites.

120
00:11:01,160 --> 00:11:06,069
And so they've successfully produced this offspring. And as we trace backwards in time.

121
00:11:06,070 --> 00:11:13,180
So this is my present day population and we trace backwards in time and we see we get quite a complicated network of relationships developing.

122
00:11:13,870 --> 00:11:18,249
And in particular it's already the case that in this generation, this individual,

123
00:11:18,250 --> 00:11:24,819
this individual and this individual all have to stay in the family tree in the sense of the pedigree of Charles,

124
00:11:24,820 --> 00:11:32,200
the second of all of the individuals down here. So all five of these individuals are in some sense descended from these three guys.

125
00:11:32,560 --> 00:11:35,680
This one left no offspring, so no one in the current generation.

126
00:11:36,010 --> 00:11:41,980
And it took me a long time to adjust the picture so that this guy doesn't actually is not actually ancestor to everybody.

127
00:11:42,010 --> 00:11:44,980
There's one person this out little exercise for you to work out which one.

128
00:11:46,860 --> 00:11:50,370
In fact, the reason that it took me so long to work it out is it's actually very difficult.

129
00:11:50,380 --> 00:12:01,050
So if this were genuinely done at random, instead of me just picking individuals after logarithm log, two of the number of individuals here,

130
00:12:01,530 --> 00:12:03,120
that number of generations back,

131
00:12:03,120 --> 00:12:11,280
we expect to see an individual who's ancestral to everybody after 1.7 log to end generations with very small variation around that.

132
00:12:11,520 --> 00:12:19,470
For large populations, everybody in the ancestral population is either ancestral to nobody now or ancestral to everybody now.

133
00:12:20,310 --> 00:12:23,640
So this suggests actually that this is not the best way to view ancestry.

134
00:12:23,820 --> 00:12:30,030
It's two joined up. So let me just actually convince you that that guy there was ancestral to everyone.

135
00:12:30,030 --> 00:12:33,749
And not only is ancestral to everyone, I've just coloured in cyan.

136
00:12:33,750 --> 00:12:37,680
I believe this colour is called all the individuals who are descended from him as I come

137
00:12:37,680 --> 00:12:42,419
down the tree without paying any attention as to whether he transmits genetic material.

138
00:12:42,420 --> 00:12:46,110
I've just looked at his offspring or her offspring since this is now.

139
00:12:47,040 --> 00:12:54,720
And you'll see that not only is everybody descended from this person, but there are multiple routes through this graph which get me from here to here.

140
00:12:55,260 --> 00:13:00,180
Okay. On the other hand, if I start thinking about genetics, you may wonder why there are small circles.

141
00:13:00,630 --> 00:13:06,780
The small circles I'm thinking of as individual genes. And if I just trace back the individual genes in this generation,

142
00:13:07,590 --> 00:13:12,600
then what I've done is I've used red lines to indicate where genes were inherited from.

143
00:13:12,600 --> 00:13:18,300
So each chooses not only an individual for a parent, but actually this gene chooses one of those two circles.

144
00:13:18,840 --> 00:13:22,320
Then this character, who is ancestral to everybody in the sense of pedigrees,

145
00:13:22,710 --> 00:13:26,670
has not transmitted any of their own genetic material to the current generation.

146
00:13:27,840 --> 00:13:35,010
Moreover, as I go backwards in time, because each gene chooses a unique parent in the previous generation,

147
00:13:35,490 --> 00:13:38,910
the structure that I get by only looking at genes is going to be much simpler

148
00:13:39,450 --> 00:13:43,350
because the number of lines that I trace as I go back in time can only get smaller.

149
00:13:43,380 --> 00:13:46,340
It's not getting bigger. I'm not having to look at two parents for a single gene.

150
00:13:46,530 --> 00:13:50,999
Single gene only has one parent, and that's easier to see if I just focus on one.

151
00:13:51,000 --> 00:13:58,590
So what I've done is I've just arbitrarily decided I'll only trace the left hand gene from each of the individuals in the present day population.

152
00:13:59,630 --> 00:14:03,650
And then you can see that as I trace backwards in time, there's a unique path that connects, for example,

153
00:14:03,650 --> 00:14:12,110
this guy to the ancestral generation, but occasionally two paths will meet and thereafter they will be the same.

154
00:14:12,530 --> 00:14:19,129
And we can encode that information in a very simple picture like this. And so what we're going to do is instead of following the haploid individual

155
00:14:19,130 --> 00:14:21,890
and trying to think about where everybody came from with both of the sorry,

156
00:14:21,950 --> 00:14:25,579
the diploid individuals and think about where everybody came from with both of their genes,

157
00:14:25,580 --> 00:14:28,580
we're just going to trace Gene by gene and see where individuals came from.

158
00:14:29,550 --> 00:14:32,700
And that leads us to the simplest imaginable model of inheritance.

159
00:14:33,120 --> 00:14:34,680
And this is called the right fisher model.

160
00:14:35,520 --> 00:14:40,920
I don't think they'd like their names being tied together like this, but I'm afraid they are inextricably connected in population genetics.

161
00:14:41,460 --> 00:14:45,300
And the Roy Fisher model is probably the most important model in mathematical population genetics.

162
00:14:45,300 --> 00:14:53,700
So it's a bit disturbing how simple it is. So the idea is that each individual chooses one parent in the previous generation.

163
00:14:53,940 --> 00:15:01,710
And I'm thinking of an individual now as being a gene. So each gene chooses one parent in the previous generation and generations discrete.

164
00:15:02,040 --> 00:15:05,370
And what happens in one generation doesn't affect what happens in the next generation.

165
00:15:07,770 --> 00:15:15,360
I told you before that in a pedigree we'd expect a common ancestor to the population on the order of long to end generations ago.

166
00:15:15,990 --> 00:15:20,010
So let's have a quick think about how far we have to trace back to get a common genetic ancestor.

167
00:15:21,040 --> 00:15:26,620
So let's take a sample size two from this population and think about how long do we have to go back before we get a common genetic ancestor?

168
00:15:27,070 --> 00:15:34,690
Well, the probability that my two individuals had a common parent is just one over ten because the first chooses a parent just uniformly at random,

169
00:15:34,960 --> 00:15:38,970
and the second one's got to choose the same parent. And the chances of choosing the same parent is one over ten.

170
00:15:39,870 --> 00:15:44,520
If they didn't choose the same parent, the chance they chose the same grandparent is one over ten.

171
00:15:45,330 --> 00:15:51,870
So the number of generations I wait until my sample of size two has a common parent's common parental

172
00:15:51,870 --> 00:15:56,970
gene is the same as the number of times I have to roll an enzyme to die before I get an end,

173
00:15:57,630 --> 00:16:05,760
and that's going to be on the order of generations. So whereas Pedigree had common ancestry after this very small number of two generations.

174
00:16:06,150 --> 00:16:10,200
Genetics genetic ancestry is determined over much longer timescales.

175
00:16:10,470 --> 00:16:13,620
And for just a pair of individuals, it would be about and generations ago.

176
00:16:13,980 --> 00:16:17,040
And in fact, for the whole population, it'll be about to end generations ago.

177
00:16:17,220 --> 00:16:24,390
So not very much longer. Okay. Now, the models we're interested in, as you may already have spotted, are very, very crude.

178
00:16:24,780 --> 00:16:29,070
They're trying to capture just some caricature of the way that populations reproduce.

179
00:16:29,100 --> 00:16:35,210
We're not going to go into fine detail of what's happening locally. So let us suppose that population size is very big.

180
00:16:35,220 --> 00:16:37,200
Otherwise you wouldn't use this kind of a model.

181
00:16:38,180 --> 00:16:43,159
So he said that for a sample of size two, it's going to take on the order of generations before we see a common ancestor.

182
00:16:43,160 --> 00:16:50,660
So let's use MN in generations as our unit of time so that the time for two individuals to find a common ancestor is of order.

183
00:16:50,660 --> 00:16:53,899
One Okay, well, if I do that,

184
00:16:53,900 --> 00:16:59,059
then the time to the most recent common ancestor is one over ten times the number of rows of my inside a die

185
00:16:59,060 --> 00:17:05,270
until I guess an N and it's well known that that is very well approximated by something completely independent,

186
00:17:05,270 --> 00:17:08,330
often called an exponential random variable with parameter one.

187
00:17:08,990 --> 00:17:13,820
So as long as I measure time in these ludicrously big, big units of generations,

188
00:17:14,030 --> 00:17:20,480
the times the most recent common ancestor of a sample of size two is on the order of one, and it's given by an exponential one random variable.

189
00:17:21,290 --> 00:17:24,469
Now let's take a bigger sample to be a bit boring. So let's take a sample of size.

190
00:17:24,470 --> 00:17:32,780
KP And Regal two three. The probability that at least three individuals in my sample have a common parent is water one over n squared.

191
00:17:32,780 --> 00:17:36,829
I claim because the first one chooses a parent, the second one's got to choose the same parent.

192
00:17:36,830 --> 00:17:39,920
That's one over in the third one's going to choose the same parent again.

193
00:17:40,010 --> 00:17:41,420
So that's one over n squared.

194
00:17:43,010 --> 00:17:49,040
I'm not going to see that happen because it's going to take me about ten squared generations before I see an event like that.

195
00:17:49,310 --> 00:17:52,400
And by then, all my lineages will have coalesced pairwise.

196
00:17:52,940 --> 00:17:57,990
So I'm never going to see three lineages coming together in a single generation in the same way.

197
00:17:58,010 --> 00:18:00,469
I'm also never going to see what we call simultaneous mergers,

198
00:18:00,470 --> 00:18:06,410
where two distinct pairs of individuals come together in the same generation because the probability this pair comes together is one over.

199
00:18:06,410 --> 00:18:11,300
In probability, this pair comes together as one over ten. So I'd have to wait and scratch generations to see it.

200
00:18:11,510 --> 00:18:17,420
And that's too long. All my lineages will have coalesced by pairwise coalescence by the time that happens.

201
00:18:18,210 --> 00:18:23,300
And so what we're left with is it's an observation of Kingman in 1982, really,

202
00:18:23,300 --> 00:18:27,470
although he proved it, lots of people had observed it was if I have a sample of size.

203
00:18:27,470 --> 00:18:30,740
K So here's a sample of size four from my population.

204
00:18:31,220 --> 00:18:33,770
Measuring time in these units of ten generations.

205
00:18:34,400 --> 00:18:39,830
The time that I have to wait as I trace backwards in time before anything happens in my genealogical trees.

206
00:18:39,830 --> 00:18:48,799
In these trees, telling me how individuals related to each other. It's just the minimum of the four choose to just reduce three exponential one

207
00:18:48,800 --> 00:18:52,820
random variables that tell me when they find their common ancestry pairwise.

208
00:18:53,540 --> 00:18:58,250
And the minimum of those two exponential random variables is just an exponential random variable.

209
00:18:58,460 --> 00:19:02,000
And we just denoted it here. Now I've got three lineages left.

210
00:19:02,210 --> 00:19:06,230
And the extra time I must wait before the next thing happens is the minimum of three.

211
00:19:06,230 --> 00:19:10,480
Choose two exponential random variables and so on. Okay.

212
00:19:10,810 --> 00:19:17,490
And here this is a picture of what this related list looks like for a sample size a thousand.

213
00:19:17,500 --> 00:19:20,530
I'm grateful to Bob Griffis for producing this for me many years ago.

214
00:19:21,670 --> 00:19:25,960
It's as you can see. A lot of stuff happens very, very quickly.

215
00:19:26,470 --> 00:19:31,060
But then we're down to a rather small number of lineages, and it takes a long, long time.

216
00:19:31,360 --> 00:19:35,110
After after this initial flurry of activity, before very much happens.

217
00:19:35,740 --> 00:19:45,580
Okay. So we've got a forwards in time model for the low frequencies and a corresponding backwards in time model given by Carmen's coalescence.

218
00:19:46,150 --> 00:19:53,530
So how does it do with data? Well, you probably guessed that this isn't really a very good model of how populations really reproduce.

219
00:19:53,770 --> 00:20:00,100
But you might hope that you could reproduce it in a laboratory. And in the 1950s, Furey tried just that.

220
00:20:00,430 --> 00:20:03,549
So what he did was he took a population of fruit flies.

221
00:20:03,550 --> 00:20:11,710
This is just awful. Melanogaster And he took the fruit flies in two different forms that differed just very slightly in their eye colour.

222
00:20:12,100 --> 00:20:16,990
So half of them, as when he started out, carried a gene which just slightly changes the eye colour.

223
00:20:17,710 --> 00:20:22,090
And he took 100 populations each consisting of eight males and females,

224
00:20:22,690 --> 00:20:26,590
and each started with half with one eye colour and half of the other eye colour.

225
00:20:27,190 --> 00:20:30,430
And he propagated these populations for 20 generations.

226
00:20:31,210 --> 00:20:37,300
And he compared the results that he got to the predictions of the Wright Fisher model.

227
00:20:37,900 --> 00:20:41,770
Now, how did he do it? He actually had to keep the populations of mice constant.

228
00:20:41,770 --> 00:20:48,370
So in each generation, he resample to always keep the constant, the population size at 16 individuals for each of those populations.

229
00:20:48,370 --> 00:20:53,800
So it must have been it doesn't seem a very big experiment by these modern standards of big data,

230
00:20:54,100 --> 00:20:57,100
but it must have been quite a tedious experiment to perform.

231
00:20:57,760 --> 00:21:05,430
And here's his results. So that right Fisher model tells us that on average, actually, the proportion of the eye,

232
00:21:05,800 --> 00:21:08,890
different flavours of AI in our population is not going to change.

233
00:21:09,130 --> 00:21:12,760
But there'll be some variability in that and it gives us a prediction for the variability.

234
00:21:13,330 --> 00:21:22,180
And he plotted So what this is one minus one, minus one over into the number of generations is what the picture should predict.

235
00:21:22,570 --> 00:21:27,700
So he plotted, all right. Plotted his results. So here are the results of his experiments.

236
00:21:27,700 --> 00:21:30,790
And this is a variance that we're plotting a variance against generation.

237
00:21:31,180 --> 00:21:35,559
So in population at the beginning, we started with exactly a half a half in all the population.

238
00:21:35,560 --> 00:21:40,600
So there was no variability. And this is just saying something about how the populations vary as time goes on.

239
00:21:40,930 --> 00:21:44,830
And eventually all the populations will be either one eye colour or the other eye colour.

240
00:21:45,160 --> 00:21:50,230
And at that point, this variance will hit point 25. So you can see it's rising steadily.

241
00:21:50,920 --> 00:21:55,720
Now, I'd like to tell you that this straight line was the prediction of the right fashion model.

242
00:21:56,440 --> 00:22:03,250
But that wouldn't be completely honest. This line, the dotted line is the prediction of the right fisher model.

243
00:22:04,230 --> 00:22:09,640
And it turns out, though, that this line is almost the right fish model.

244
00:22:09,660 --> 00:22:16,530
But instead of taking the true population size, which in this case is 16, I've substituted 11 and a half.

245
00:22:17,630 --> 00:22:22,960
And by virtue of doing that, I mean, I know it's not a perfect fit, but actually for an experiment of this size, that's pretty good.

246
00:22:23,930 --> 00:22:25,430
And it turns out that's universal.

247
00:22:26,180 --> 00:22:33,050
It is pretty good as long as I don't use the real population size and I put in a population size to suit my purposes.

248
00:22:33,590 --> 00:22:38,270
The Kingman coalescence, or the right Fisher model is a pretty good approximation, even to natural populations.

249
00:22:38,480 --> 00:22:40,670
As long as I sample individuals from far away,

250
00:22:41,180 --> 00:22:47,360
far enough away from one another that I'm not seeing local effects brought about by them living in very close proximity, for example.

251
00:22:48,170 --> 00:22:51,830
And it's an example of the sort of scale of this fudge factor.

252
00:22:52,160 --> 00:22:55,610
So how much do I have to change the population size to make things fit?

253
00:22:56,030 --> 00:23:03,409
I think the human population is rather nice. So for the whole world I would need to take an effective population size.

254
00:23:03,410 --> 00:23:07,100
I'd need a substitute and to be about 50,000 in my right vision model.

255
00:23:07,910 --> 00:23:11,720
And of course, the true population size of humans is 7 billion.

256
00:23:12,590 --> 00:23:21,890
So the difference between the number I plug in to make my model fit and the true number is not five orders of magnitude.

257
00:23:22,550 --> 00:23:25,670
So that seems crazy that this should work, is completely mad that it should work.

258
00:23:25,670 --> 00:23:30,200
But actually beyond that correction, it fits the data extremely well.

259
00:23:31,810 --> 00:23:35,020
Okay. Now, we would like to understand why.

260
00:23:35,500 --> 00:23:39,159
And in particular, we would like to understand if we're going to understand our basic question,

261
00:23:39,160 --> 00:23:44,410
how the different forces of evolution feed into this need to make it all work so well.

262
00:23:44,650 --> 00:23:48,190
What the right fish model is doing is it's just modelling the genetic drift.

263
00:23:48,520 --> 00:23:52,600
But how would selection alter any how the spatial structure of any.

264
00:23:53,690 --> 00:24:00,489
And I started working on this stuff about 20 years ago when Nick Boulton, who is a very distinguished evolutionary geneticist now,

265
00:24:00,490 --> 00:24:09,640
at least in Australia, came to see me and Nick said, Well, look, I'm studying these grasshoppers who they are and they live in the maritime Alps.

266
00:24:09,730 --> 00:24:13,270
And you'll see, like he always chooses nice mountain ranges for his field trips.

267
00:24:13,780 --> 00:24:16,839
And they they really are in a spatial continuum.

268
00:24:16,840 --> 00:24:22,060
And I want to know how this spatial structure is affecting the genetics of pedestrian pedestrians.

269
00:24:22,780 --> 00:24:28,780
And I should tell you why it's called disruptive industries, because it is a pedestrian grasshopper.

270
00:24:28,870 --> 00:24:33,280
It's hard to see, but this is the vestigial wing. This thing cannot fly.

271
00:24:33,670 --> 00:24:37,659
Okay. So it crawls around, hops around. It doesn't move very far in its lifetime.

272
00:24:37,660 --> 00:24:41,920
So the spatial structure to as a pedestrian probably looks pretty much like the plane.

273
00:24:42,590 --> 00:24:48,100
And at the same time, Nick said, oh, and by the way. Right. And Marco almost solved this in the 1940.

274
00:24:48,110 --> 00:24:52,450
So Gustav Malago is another of the greats of population genetics.

275
00:24:53,080 --> 00:24:57,580
And the way that right in Monaco solved it was they took the right fishing model and they adapted it to a spatial setting.

276
00:24:57,580 --> 00:25:00,700
So they said, Let's suppose individuals are scattered across space.

277
00:25:00,700 --> 00:25:07,269
This pointer is really not good. So we go so here individuals scattered across space and it's in a kind of

278
00:25:07,270 --> 00:25:10,570
uniform way that each of them just chooses where they fall uniformly at random.

279
00:25:10,990 --> 00:25:15,060
And this has been drawn on a source for reasons that will become clear in due course.

280
00:25:15,400 --> 00:25:21,640
And I will show some realisations of a simulation due to Jerome Kelleher, who I think is also sitting up there somewhere.

281
00:25:22,610 --> 00:25:25,399
And the way that their model works is it evolves in discrete generations,

282
00:25:25,400 --> 00:25:30,889
just like the Wright Fisher model and the number of offspring that each individual

283
00:25:30,890 --> 00:25:35,930
produces in each generation is taken to be a Poisson random variable with parameter one.

284
00:25:36,290 --> 00:25:39,530
So why is that what they chose when they chose that? Because in the right fisher model,

285
00:25:39,830 --> 00:25:46,340
if you look at the number of number of offspring that a single individual produces phylogeny, it's approximately a Poisson.

286
00:25:46,370 --> 00:25:53,520
It's very, very close to being just a partial. So now here, Mitch Gooding drew this for me.

287
00:25:53,810 --> 00:25:58,550
Here's a histogram which just tells you this is how many times I should expect to get zero offspring.

288
00:25:58,820 --> 00:26:04,040
This is how many times I should expect to get one offspring. So quite a lot of the time, and this is two offspring and so on.

289
00:26:04,040 --> 00:26:07,610
But on the average I produce one offspring. Okay.

290
00:26:08,150 --> 00:26:11,510
And I can't have my offspring all sitting on top of each other. That wouldn't be right.

291
00:26:11,840 --> 00:26:16,640
And so I scatter them around the position of the parents according to a Gaussian distribution.

292
00:26:16,820 --> 00:26:20,150
So they're just distributed close by in a nice, symmetric way.

293
00:26:20,750 --> 00:26:28,640
So understandably. Right. And Marco thought that this sort of pattern would persist, that if they looked in their population in generation ten,

294
00:26:28,970 --> 00:26:32,930
it would still look quite a lot like this one, still look pretty uniformly spread.

295
00:26:34,070 --> 00:26:40,790
I'm working on that assumption. They were able to do what was the equivalent in the 1940s of writing down those genealogical trees,

296
00:26:40,790 --> 00:26:45,049
telling us how individuals in the population were related to each other and genetically,

297
00:26:45,050 --> 00:26:49,480
how that the correlations between the genetic type would decay with distance.

298
00:26:49,490 --> 00:26:59,410
They predicted it would decay approximately exponentially. But then in 1975, Joe Feldstein noticed that actually the assumptions were inconsistent.

299
00:27:00,070 --> 00:27:03,729
And this is during the simulation, which he's probably rather embarrassed I'm using.

300
00:27:03,730 --> 00:27:09,100
So he did it for a lab meeting a long time ago. But here's the initial condition and what he's done.

301
00:27:09,100 --> 00:27:14,710
He's working on a tourist, and he's suppose that the population really does evolve according to the Malago model.

302
00:27:15,250 --> 00:27:19,990
And after ten generations, this is what it looks like. So the population is still pretty much a thousand.

303
00:27:20,260 --> 00:27:24,370
It's not changed very much, but we're getting these white spaces developing.

304
00:27:25,550 --> 00:27:29,630
After a hundred generations, the population is still pretty close to a thousand, actually.

305
00:27:29,960 --> 00:27:33,290
But we really are getting a lot of white space. The population is really clamping.

306
00:27:34,860 --> 00:27:40,950
By a thousand generations, the population is caving in and realising that mathematics cannot be defeated because there is a

307
00:27:40,950 --> 00:27:46,320
theorem that says it has got to die out eventually and it's noticing that it really ought to.

308
00:27:46,350 --> 00:27:51,450
So it's down to close to 300, but those individuals that are left are really clustered together.

309
00:27:51,780 --> 00:27:59,459
Okay. Now, a fellow Sunstein first observed this when he was working on the whole if the real plane hole of Euclidean space and he said,

310
00:27:59,460 --> 00:28:03,480
oh, well, maybe it's just cause I'm working with an infinite population. Real populations are finite.

311
00:28:03,490 --> 00:28:04,979
Let's look at a finite population.

312
00:28:04,980 --> 00:28:13,410
So he moved on to a Taurus and then he noticed, well, unfortunately, on a tourist, either the population blows up or it dies out, so it will die out.

313
00:28:13,410 --> 00:28:18,330
If I just have on average one offspring and if I slightly increase that, the population will just explode eventually.

314
00:28:18,340 --> 00:28:26,430
So that doesn't work. And then he said, Well, let's suppose that somehow the total population size, oh, my Taurus is exogenous.

315
00:28:26,430 --> 00:28:28,500
And he specified and had to fix it to be a thousand.

316
00:28:29,460 --> 00:28:35,580
But we can see from the simulation that that's still not working, actually, because the population size here didn't change very much.

317
00:28:35,730 --> 00:28:37,260
It's still pretty close to a thousand,

318
00:28:37,500 --> 00:28:44,190
but we're still getting clumping and Feldstein realised that this was going to happen and at that point he said, okay, I give up.

319
00:28:44,550 --> 00:28:48,270
And he wrote a paper which famously dubbed this problem the pain in the Taurus.

320
00:28:49,170 --> 00:28:53,130
And so the challenge that Nick was really throwing down to me was to solve the pain in the Taurus.

321
00:28:54,430 --> 00:29:01,639
All right. We want you to produce a model which was a little bit like the Moloko assumption

322
00:29:01,640 --> 00:29:05,090
that the population would be distributed in space in a relatively uniform manner,

323
00:29:05,390 --> 00:29:10,100
but which actually had a stability to it so that we could write down genealogical trees in a consistent way.

324
00:29:11,200 --> 00:29:13,660
And we want that model to address one or two other issues.

325
00:29:14,200 --> 00:29:19,320
So we've already seen that genetic diversity is much, much lower than what you'd expect from census numbers.

326
00:29:19,360 --> 00:29:25,450
That's another way of saying the effect of population size is orders of magnitude different from the census population size.

327
00:29:25,460 --> 00:29:30,220
That's that same statement. In other words, another thing was I said that right.

328
00:29:30,220 --> 00:29:36,490
And Malika observed that the correlations between genetic types would decay, sort of expand exponentially with distance apart.

329
00:29:37,530 --> 00:29:42,180
That's sort of true ish over some scales. But then when you look over larger scales,

330
00:29:42,420 --> 00:29:47,100
the rate of exponential decay appears to have decreased and you get longer range correlations than seems reasonable.

331
00:29:48,190 --> 00:29:55,030
And a possible explanation for this is that the demographic history of many populations is really dominated by large scale events.

332
00:29:55,060 --> 00:29:58,970
So imagine I'm a population of plants living on a forest floor.

333
00:29:59,470 --> 00:30:04,120
Every hundred generations or so I forest fire sweeps through and completely wipes the population out.

334
00:30:04,450 --> 00:30:07,630
And then it gets very, very rapidly replaced, re colonised.

335
00:30:08,110 --> 00:30:15,910
And that's going to lead to both a reduction in genetic diversity and long large scale correlations in LEO frequencies.

336
00:30:16,360 --> 00:30:20,520
Now I am not claiming the model I am about to write down. I think it is quite a good model for forest fires.

337
00:30:20,530 --> 00:30:24,160
It's not a very good model for glacial maximum for ice ages.

338
00:30:24,520 --> 00:30:31,720
But I like to show this slide just to remind me, to give you a notion of the sort of timescales over which evolution is happening.

339
00:30:32,440 --> 00:30:35,979
So remember, we set the effect of population size for the human population.

340
00:30:35,980 --> 00:30:39,400
Let's say if we just look at European humans is about 20,000.

341
00:30:40,240 --> 00:30:47,440
So that means that my genetic composition is being determined over 20,000 generations and into generation time.

342
00:30:47,830 --> 00:30:54,940
I don't know. We could argue over 20 years. So that means we're talking about timescales of hundreds of thousands of years.

343
00:30:55,660 --> 00:31:02,620
Right. The last ice age, the last glacial maximum Northern Europe was largely covered in ice.

344
00:31:03,310 --> 00:31:06,910
Humans did not live there, and that was only 18,000 years ago.

345
00:31:07,150 --> 00:31:09,820
So from the point of view of genetics, that's just the twinkling of an eye.

346
00:31:10,570 --> 00:31:17,760
So these large scale events are really going to have influenced our genetic composition, and we can't simply ignore them.

347
00:31:17,770 --> 00:31:20,660
They really are going to have affected things. Okay.

348
00:31:20,960 --> 00:31:28,970
Another thing I wanted to say before showing you the model that we came up with is one we derived derived the Cayman coalescence.

349
00:31:29,390 --> 00:31:35,000
I emphasised that for large populations we'd never see three lineages merging in

350
00:31:35,000 --> 00:31:38,150
a single generation because of having a common parent in a single generation,

351
00:31:38,480 --> 00:31:41,180
because individuals had so many parents to choose from,

352
00:31:41,180 --> 00:31:47,180
you never got three of them choosing the same ones on the timescale that we were looking at, where you just get pairs choosing the same one.

353
00:31:47,510 --> 00:31:49,820
And that's just because in squared was much bigger than in.

354
00:31:50,900 --> 00:31:56,630
In a spatial continuum if I'm an individual and I'm looking around me for potential parents in the previous generation.

355
00:31:57,690 --> 00:32:03,690
That may not be true and may not be very large. And it may be the case that any squared is really not so much bigger than N.

356
00:32:03,690 --> 00:32:10,140
And so I will see mergers of not just two ancestral lineages at a time, but three, four or five, any number.

357
00:32:11,180 --> 00:32:14,600
Okay. So here's his model. Our first first stab at a model.

358
00:32:15,410 --> 00:32:19,040
We going to suppose the population is just spread out in space in the same way as Wright Malachi did.

359
00:32:19,970 --> 00:32:24,340
But no reproduction isn't going to be based on individuals. It's going to be based on events.

360
00:32:24,380 --> 00:32:29,150
And this is the key insight. So what we do is we throw events down.

361
00:32:29,300 --> 00:32:34,640
You could do this in discrete generations. Mathematically, it's convenient to do it in overlapping generations, as we say.

362
00:32:34,650 --> 00:32:43,410
So we throw down one event at a time. And a reproduction event is just going to affect a region which is determined by the nature of the event.

363
00:32:43,430 --> 00:32:46,129
So here is just a disk sent x and radius.

364
00:32:46,130 --> 00:32:51,920
Ah, now if the region I throw down is empty, I don't do anything because there's no population there to reproduce.

365
00:32:52,430 --> 00:32:58,760
But if it's not empty first, among all the individuals living there, I choose a parent uniformly at random.

366
00:32:59,440 --> 00:33:05,810
The first thing to notice is if I'm living in a very crowded region, the odds of being chosen as a parent get to be very small.

367
00:33:06,410 --> 00:33:10,630
So my reproductive success, if I live in a very crowded region, is not big.

368
00:33:10,640 --> 00:33:15,410
It's very small. On the other hand, if I'm in a very sparsely populated region, I will be picked.

369
00:33:15,810 --> 00:33:21,950
Okay. And that's the key. That's what prevents that clumping that we saw from Feldstein's observations.

370
00:33:22,520 --> 00:33:26,599
Okay, so I've chosen this individual. I'm going to kill a proportion of the population.

371
00:33:26,600 --> 00:33:29,630
In this example, I allowed the parent to die. I don't have to.

372
00:33:30,760 --> 00:33:35,469
And I replace them. So they are killed with some probability that I just specify independently.

373
00:33:35,470 --> 00:33:39,970
So each of them flipped a coin that comes off heads with probably see you. And if it comes up heads, they die.

374
00:33:41,070 --> 00:33:46,790
And then I replace them with offspring and they're scattered in the same way as the parental population is scattered

375
00:33:46,800 --> 00:33:52,740
just by picking points uniformly at random in the region and the distribution of the number of offspring.

376
00:33:52,890 --> 00:33:55,980
It's random, but it's chosen to roughly replenish the population.

377
00:33:55,980 --> 00:33:59,960
So the population density should be, roughly speaking, constant. Okay.

378
00:33:59,970 --> 00:34:04,140
So and then we remove the dead individuals, and that's my new population.

379
00:34:05,100 --> 00:34:09,300
Okay. So how does this work as a model? It's obviously very crude. It's a little bit like a white fisher model.

380
00:34:09,930 --> 00:34:13,140
Does it work at all? Well, it overcomes the pain in the tourists.

381
00:34:13,170 --> 00:34:17,730
It does have a nice stationary distribution with populations distributed uniformly across space.

382
00:34:18,710 --> 00:34:22,010
It allows us to incorporate large scale extinction recolonisation events very

383
00:34:22,010 --> 00:34:26,180
easily and it also is easy to extend to include things like natural selection.

384
00:34:26,180 --> 00:34:30,169
So for example, I might select the parent not just uniformly among those in the region,

385
00:34:30,170 --> 00:34:33,790
but according to their genetic type weighted by their genetic type.

386
00:34:33,800 --> 00:34:37,430
Or I might choose individuals to die according to their genetic type.

387
00:34:38,030 --> 00:34:44,719
So it's very easy to adapting to natural selection and we can write down the distribution of those genealogical trees,

388
00:34:44,720 --> 00:34:47,840
the things that the geneticist transferring. The problem is.

389
00:34:49,450 --> 00:34:56,139
It's a bit of a mess. So the expressions we write down are extremely complicated, but it's only a metaphor.

390
00:34:56,140 --> 00:35:03,790
Mathematical reasons, at least as the population is relatively dense, that mathematical mess can all be approximated by a single model.

391
00:35:04,600 --> 00:35:09,610
And the way that we think about that single model is that what we're going to approximate

392
00:35:09,610 --> 00:35:12,910
is what we're going to use as an approximate in model as a model for sampling probability.

393
00:35:12,920 --> 00:35:19,120
So what my model answers the question if I were to sample an individual from the point X at time T,

394
00:35:19,870 --> 00:35:24,640
what is the probability that it is a type A, whatever type A might be?

395
00:35:25,120 --> 00:35:30,139
So that's the question that our model would let us answer. And to explain how it works.

396
00:35:30,140 --> 00:35:35,720
It's convenient just to forget space for a moment because we're just going to adapt a norm spatial model to a spatial model.

397
00:35:36,350 --> 00:35:41,420
So here's how it's going to work. Reproduction is, again, going to be based based on events exactly as it was before.

398
00:35:41,750 --> 00:35:44,899
So events now specified a time.

399
00:35:44,900 --> 00:35:47,480
That's the time when the event happens and an impact.

400
00:35:47,660 --> 00:35:51,770
This you that's the proportion of the population that's going to be affected by the reproduction event.

401
00:35:52,560 --> 00:35:57,230
And let's have a look at this event and see what it does to our population so immediately before the event.

402
00:35:57,680 --> 00:35:59,329
This is how the different types are distributed.

403
00:35:59,330 --> 00:36:03,830
I've just used three different colours to represent three different types, and I got to select a parent.

404
00:36:03,980 --> 00:36:08,670
So I'm going to select my parent uniformly at random from the population. So I just threw a point just at random.

405
00:36:08,670 --> 00:36:12,140
I'm zero one here. And it happened to land here in this cyan region.

406
00:36:12,680 --> 00:36:19,520
So the type of the parent is going to be cyan. Now a proportion of my population is going to be killed.

407
00:36:20,030 --> 00:36:24,859
The remaining one minus you survive. So this band here is one minus two times the width of that one.

408
00:36:24,860 --> 00:36:28,700
That's one minus you times that one. And this bit is one minus you times that one.

409
00:36:29,650 --> 00:36:33,910
And I replace everyone I killed with offspring of this chosen type.

410
00:36:34,570 --> 00:36:42,090
And so on. This example, they're all saying. And the nice thing about this model is it's very easy to write down how the ancestral lineages behave.

411
00:36:42,870 --> 00:36:49,410
So here I've taken a sample from my population. The sample is a size five and I'm wondering how it's going to behave.

412
00:36:49,420 --> 00:36:52,110
So what's going to happen is I traced backwards through this event.

413
00:36:52,980 --> 00:36:58,530
Well, as it happened, these two guys fell in the region of the population that survived the event.

414
00:36:58,860 --> 00:37:02,250
And so they just survive. Nothing happens to them. They're still distinct lineages.

415
00:37:02,430 --> 00:37:08,830
And the previous generation. But these three all fell in the portion of the population corresponding to offspring.

416
00:37:09,490 --> 00:37:11,220
And so we know they had a common parent.

417
00:37:11,230 --> 00:37:17,710
So here we have an example of not a pairwise merger, but a three merger, and they merge into this common ancestral lineage.

418
00:37:18,010 --> 00:37:23,260
And it's very, very easy to write down mathematical expressions for the probabilities of events like this.

419
00:37:24,480 --> 00:37:29,040
Okay. So the idea of our approximation to the model,

420
00:37:29,070 --> 00:37:35,190
as I wrote down just now and it can be obtained genuinely is the limit of that model is that we do the same thing in space.

421
00:37:35,880 --> 00:37:42,990
So now we're not just specifying the distribution of types in a single region as we were in a non spatial example just now.

422
00:37:43,380 --> 00:37:51,810
But for each point Z And each time T I'm telling you the distribution of the type of an individual sample from the population at Z at time t.

423
00:37:51,990 --> 00:37:56,730
So if I sample an individual at time t. From this point, what's the probability it's type A?

424
00:37:56,760 --> 00:38:00,659
That's the question I'm answering. Okay. And it's much the same as what we did before.

425
00:38:00,660 --> 00:38:04,770
Reproduction events affect bounded regions. Those regions are now never empty.

426
00:38:04,770 --> 00:38:08,499
So I don't need to worry about empty space. I got to sample of parents.

427
00:38:08,500 --> 00:38:12,000
So first I'm going to sample a location from the parent. So this point Z was just uniform.

428
00:38:13,050 --> 00:38:18,360
Then I choose a type according to the distribution there, and it came out to be read much as it did on the previous slide.

429
00:38:19,410 --> 00:38:27,160
And then I update for everybody in this region. I kill a proportion of individuals and I replace them by offspring of this chosen type.

430
00:38:27,660 --> 00:38:34,229
And this is the fancy mathematical way of writing it. But all it's saying is that everywhere in this region I delete just slice off

431
00:38:34,230 --> 00:38:37,800
a proportion of the population and replace it by individuals of this type.

432
00:38:39,660 --> 00:38:43,940
I can write down a backwards in time model that corresponds to that that tells me about these genealogical trees.

433
00:38:44,750 --> 00:38:51,230
Because if I'm just a single individual in a sample, I want to know how the ancestry of the individual in my sample evolves backwards in time.

434
00:38:51,770 --> 00:38:56,960
I wait until the first time my individual is in a region that is affected by one of these reproduction events,

435
00:38:57,680 --> 00:39:00,890
and then it's got a probability you have been offspring of the event. Right.

436
00:39:01,040 --> 00:39:04,820
And if it is, then it's going to have to jump to the location of the parent.

437
00:39:05,030 --> 00:39:08,930
If it's not, nothing happens. It just keeps going. And what's the location of the parent?

438
00:39:08,960 --> 00:39:10,700
Well, it was just sampled uniformly from the ball.

439
00:39:10,700 --> 00:39:16,250
So it's very easy mathematically to write down expressions for the way these ancestral lineages move around in time.

440
00:39:17,200 --> 00:39:23,860
And if a region happens to cover a whole collection of ancestral lineages, as it did here, the idea is these green guys are ancestral lineages.

441
00:39:24,310 --> 00:39:29,629
Lineages outside the region can't be affected, but inside. Each end of each led each flipped.

442
00:39:29,630 --> 00:39:33,410
A coin that comes up heads is probably see you if it comes up heads we.

443
00:39:33,740 --> 00:39:40,010
This was a an offspring of the event and these three offspring must all be descended from the common parent.

444
00:39:40,220 --> 00:39:42,800
And the location of the common pair was uniform across the bow.

445
00:39:43,130 --> 00:39:51,710
So individuals coalesce when they within or can coalesce when they're within region, the same region that's affected by an event.

446
00:39:52,130 --> 00:39:55,940
So that gives us a backwards and forwards in time model. Okay.

447
00:39:55,940 --> 00:39:59,480
So it looks pretty crude and you probably think it's not going to have anything to do with data.

448
00:40:00,770 --> 00:40:04,340
But remember that Kingman worked over very, very large scales.

449
00:40:05,350 --> 00:40:09,670
Even though the Kingman coalition coalition was based on this very crude white fisher model.

450
00:40:10,180 --> 00:40:12,819
So is it the case that if we look over kind of intermediate scale,

451
00:40:12,820 --> 00:40:18,820
somewhere between the scale on which Kingman works and very local scales on which this model is clearly rubbish, it might work.

452
00:40:19,690 --> 00:40:22,990
And when you look at low frequencies, you think, Oh, no way.

453
00:40:23,200 --> 00:40:29,019
So this is just what a pattern of low frequencies might look like after we thrown down 50 odd events.

454
00:40:29,020 --> 00:40:32,650
And it just doesn't look realistic. But if I look over larger scales, maybe it will.

455
00:40:34,520 --> 00:40:37,910
So here's some something to try to convince you that there might be something in it.

456
00:40:37,940 --> 00:40:42,589
So this is this is a horrible bacterium, actually, pseudomonas aeruginosa.

457
00:40:42,590 --> 00:40:45,980
It's found in the lungs of cystic fibrosis patients, amongst other things.

458
00:40:46,760 --> 00:40:50,750
And this picture is from Kevin Foster's lab in the zoology department in Oxford.

459
00:40:51,260 --> 00:40:52,760
And I find this very beautiful.

460
00:40:52,760 --> 00:41:00,980
He has an incredibly high resolution microscope which allows him to observe that you can just about pick out, I hope, these individual bacteria.

461
00:41:01,950 --> 00:41:05,670
It's just an extraordinary picture, but we're not trying to model this.

462
00:41:05,700 --> 00:41:09,810
It's obvious that what's going on is this is the edge of a bacterial colony as it's evolving.

463
00:41:10,260 --> 00:41:13,290
So this is just a snapshot of it. This is empty space and these are the bacteria.

464
00:41:13,980 --> 00:41:20,160
And obviously, what happens next depends on the particular configuration of these bacteria and their exact load shape.

465
00:41:21,370 --> 00:41:26,530
But that's not what we're trying to capture. Let's zoom out a bit and look at that big bacterial colony on a slightly larger scale.

466
00:41:27,130 --> 00:41:31,120
So this is more like visible scales. And what we see is emerging structure.

467
00:41:31,960 --> 00:41:40,090
So what Kevin's done here is he's taken two populations of the bacterium, one blue and one green, but they are equally fit.

468
00:41:40,690 --> 00:41:44,139
And the beginning of the experiment, he's just makes them all up and he's dropped,

469
00:41:44,140 --> 00:41:48,460
put a little droplet onto his nutrient plate and you can sort of still see the vestiges of it there.

470
00:41:48,820 --> 00:41:49,750
And he's allowed it to grow.

471
00:41:50,560 --> 00:41:58,090
And as it grows, we see these sectors developing and these sectors are kind of proxy for relatedness amongst individuals in the population.

472
00:41:58,870 --> 00:42:05,980
So, for example, all these individuals in this blue sector will descend from some common ancestral bacterium back here somewhere.

473
00:42:06,510 --> 00:42:11,870
Okay. So what if we do the same thing for our model? I saw two basset.

474
00:42:13,030 --> 00:42:16,990
We're obviously picking out the same basic structure, the same secretary.

475
00:42:17,020 --> 00:42:22,590
And again, this is a simulation I owe to Jerome Callahan. So we've done exactly the same thing.

476
00:42:22,600 --> 00:42:26,530
We've put a mixture down at time zero. We've allowed it to evolve fortuitously.

477
00:42:26,530 --> 00:42:34,899
We've chosen the same colours as Kevin and you see sectors of much the same sort of pattern developing and these sectors are ubiquitous.

478
00:42:34,900 --> 00:42:41,440
It's not just pseudomonas that does it. This is yeast. This is from a paper of Oscar how Jack and his co-workers from 2007.

479
00:42:42,130 --> 00:42:47,870
By changing one parameter in our model or the ratio of two parameters, we can reproduce pictures like this.

480
00:42:47,870 --> 00:42:51,459
And I'm sorry, I forgot to email Jerome and ask him for one so I don't have one to show you,

481
00:42:51,460 --> 00:42:58,450
but I assure you that we can also reproduce these narrower sectors that are characteristic of yeast.

482
00:42:58,930 --> 00:43:04,430
Okay, so with that sort of reassurance, we decided to press on and try and understand a bit more about spatial structure.

483
00:43:04,430 --> 00:43:08,740
And I wanted to tell you about some very recent work on things called hybrid zones.

484
00:43:08,830 --> 00:43:11,020
So now I have to tell you some biology and show you pretty pictures again.

485
00:43:11,920 --> 00:43:17,030
So as Nick Bolton's got older and 20 years ago, he could just about catch up with Buddhism,

486
00:43:17,080 --> 00:43:23,260
a pedestrian that crawls very slowly around the maritime Alps. He's realised there are things called plants and they hardly move at all.

487
00:43:24,190 --> 00:43:28,510
So this is again Nick, who's a good friend, so I can tell you things about him.

488
00:43:28,900 --> 00:43:32,770
This is anti rainham and these on Geranium live in the Pyrenees.

489
00:43:33,400 --> 00:43:40,570
Obviously they don't live in some sort of damp field in the West Midlands, they live in the Pyrenees and they exhibit what's called a hybrid zone.

490
00:43:40,690 --> 00:43:48,880
So a hybrid zone is when you get two genetically distinct populations coming together and at the interface between them,

491
00:43:49,120 --> 00:43:52,450
they're sufficiently similar genetically that they can reproduce and hybridise.

492
00:43:52,870 --> 00:43:57,640
But the hybrids are not as fit as the pure populations were.

493
00:43:58,180 --> 00:44:04,120
And these are ubiquitous. And if you think about plant populations in the last place, you maximum that we talked about before,

494
00:44:04,750 --> 00:44:08,559
a lot of plants were sort of pushed back into Refugia after the Ice Age,

495
00:44:08,560 --> 00:44:10,870
they started to expand again and when they came back together,

496
00:44:11,140 --> 00:44:17,830
they were sufficiently genetically distinct that you could distinguish them and that when they interbred with one another, the hybrids were less fit.

497
00:44:18,190 --> 00:44:25,210
And so this particular hybrid zone on one side of the zone, the and Geronimo Yellow, and on the other side that is some pinkish purple colour.

498
00:44:26,050 --> 00:44:26,300
Okay.

499
00:44:26,710 --> 00:44:32,980
And hybrid zones are maintained by a balance between the desire of the plants to spread their offspring out and this selection against the hybrids.

500
00:44:34,090 --> 00:44:37,360
I said they're ubiquitous. Here are a couple of textbook examples.

501
00:44:37,720 --> 00:44:43,570
So the one you see in every textbook is this one up here. This is Mice with Musketeers and most domestics.

502
00:44:44,080 --> 00:44:48,580
So in the Northeast, mice take this form if you catch one in your larder.

503
00:44:49,270 --> 00:44:54,159
And down here we have the most domestics and you can probably see much better than I

504
00:44:54,160 --> 00:45:00,100
can that there is a narrow hybrid zone in this colour between the two populations.

505
00:45:01,090 --> 00:45:06,850
Here's another one. I like this one. This is the fire belly toad against the yellow bellied toad.

506
00:45:07,450 --> 00:45:10,760
And they have a really wacky hybrid scene. So you can see the hybrids on here.

507
00:45:10,780 --> 00:45:13,560
It goes all the way. Through Europe.

508
00:45:15,470 --> 00:45:21,200
And what this picture does, what this slides down is it's focussed on this little bit of this hybrid zone where a lot of experiments have been done.

509
00:45:21,200 --> 00:45:27,169
In hybrid zone, there's almost flat. It's about 20 kilometres wide and they've plotted a low frequency.

510
00:45:27,170 --> 00:45:32,750
So the frequency is the genetic type that gives you this toad versus the frequency, the genetic type that gives you this toad.

511
00:45:33,260 --> 00:45:37,570
And these are data points that they've plotted. Now in a region where.

512
00:45:37,990 --> 00:45:41,380
Well, okay, if you if you don't believe in genetic drift.

513
00:45:41,920 --> 00:45:48,219
I do believe in genetic drift. But if you don't believe in genetic drift, then you can model these hybrid science using this differential equation.

514
00:45:48,220 --> 00:45:51,100
It's called the it's a special case of what's called the Alan Cohen equation.

515
00:45:52,520 --> 00:45:59,480
We model it with this plus noise in some sense, so plus some genetic drift term, but it's easier to write it down in this example.

516
00:46:00,660 --> 00:46:08,700
And what that predicts is that actually across the hybrid zone, this should be in a curve and the curve would look a bit like one plus one over two.

517
00:46:09,060 --> 00:46:14,520
And those of you who can remember what one plus hyperbolic tangent over two looks like will think, Gosh, actually, that's not bad.

518
00:46:15,490 --> 00:46:19,270
I mean, that's the right sort of shape, but it predicts a relatively stable hybrid scene.

519
00:46:19,810 --> 00:46:26,170
But the question we set out to ask was, well, okay, that's what it looks like now, but how is this hybrid in itself going to evolve with time?

520
00:46:27,150 --> 00:46:32,850
If you zoom out, I mean, this one's 20 kilometres wide. You don't have to be very far away before it looks like a sharp interface.

521
00:46:33,880 --> 00:46:38,320
And so with Nick Freeman, who's now a lecturer in Sheffield, and Sarah Pennington,

522
00:46:38,320 --> 00:46:43,780
who is about to take up a research fellowship in mathematics here in the Institute and at New College,

523
00:46:44,800 --> 00:46:51,790
we have shown that at least if we start from sufficiently regular initial conditions to make myself mathematically honest as we zoom out,

524
00:46:52,330 --> 00:46:58,360
the hybrid zone becomes sharp, and that's whether we use this deterministic differential equation that the guys use,

525
00:46:58,360 --> 00:47:03,160
and they already knew this result of their deterministic equation or whether we also add to noise.

526
00:47:03,170 --> 00:47:05,710
So let's understand what happens as we zoom out.

527
00:47:06,100 --> 00:47:11,200
So as we zoom out, the hybrid zoom becomes sharp, that sort of clear because it was only 20 kilometres wide in the first place.

528
00:47:11,590 --> 00:47:16,570
But how does it move? It involves a cool design called curvature flow and to understand curvature flow.

529
00:47:17,700 --> 00:47:23,390
Is a hybrid zone, a putative hybrid zone, roughly speaking, the curvature at this point.

530
00:47:23,750 --> 00:47:28,309
You take the biggest circle that you can. That just fits here, doesn't crossover.

531
00:47:28,310 --> 00:47:32,450
It just fits. And the curvature is one over the radius of that circle.

532
00:47:33,650 --> 00:47:38,150
And here the circles on the other side and the curvature is one over the radius of this circle.

533
00:47:38,420 --> 00:47:41,390
So the curvature here is less than the curvature there.

534
00:47:41,630 --> 00:47:49,310
And it also has the opposite sign and curvature flow will push this point inwards and this point outwards and see it in action.

535
00:47:50,420 --> 00:47:58,580
Mat Dunlop is a student in the University of Warwick and he very kindly produced an illustration of curvature flow for me.

536
00:47:59,000 --> 00:48:01,400
Now let me just stop it so you can see the initial condition.

537
00:48:01,760 --> 00:48:13,069
I did not choose the name of this file, so you can see the Batman symbol very quickly degenerates into a sausage on up here where it's almost flat.

538
00:48:13,070 --> 00:48:16,190
You know, nothing much happens. And these ends are pushing and pushing in.

539
00:48:18,550 --> 00:48:21,220
And then there's still not much happening.

540
00:48:21,460 --> 00:48:27,460
But what we're going to see is that these ends will eventually there's not going to be any flat bit left because these ends have pushed in so far.

541
00:48:27,880 --> 00:48:31,540
And this will become circular. In fact, any convex region would eventually become circular.

542
00:48:32,170 --> 00:48:38,320
And as this shrinks, it's going to shrink is going to go faster and faster and faster because the circle gets smaller, the coverage is getting bigger.

543
00:48:38,500 --> 00:48:41,740
So many of which are flow goes faster and it's gone. Okay.

544
00:48:42,670 --> 00:48:46,020
So that's what I mean. Curvature flow does for you. So.

545
00:48:47,930 --> 00:48:53,180
What's going to happen to our hybrid teams. So now what's going to happen if we put some noise in?

546
00:48:53,180 --> 00:48:56,630
What's going to happen if we use our spatial and a Fleming view model and puts noisy?

547
00:48:58,480 --> 00:49:04,080
Well. This is a video that Nick Freeman did.

548
00:49:05,430 --> 00:49:13,010
So you can see again, he'd started with something a little bit like a Batman and Batman symbol, but he didn't have the imagination to do that.

549
00:49:13,020 --> 00:49:16,980
And you can see, again, it's pushing out to be sort of sausage shaped, but there's a bit of noise here.

550
00:49:17,190 --> 00:49:20,430
But you could imagine if you looked at this from far enough away and had bad enough eyesight.

551
00:49:20,550 --> 00:49:28,680
So for me, this looks a lot like a bunch of flow. I'm going to speed it up a bit cause we're getting to the end.

552
00:49:29,130 --> 00:49:33,660
So let's zoom along a bit at all. And you see it really is doing what COVID flowed.

553
00:49:33,660 --> 00:49:39,480
It did for the Batman. Same symbol. It's gone almost round and.

554
00:49:40,520 --> 00:49:45,730
If I keep going, it get smaller. Okay. Now, Jerome did another one for me.

555
00:49:46,210 --> 00:49:49,570
Jerome did one that looks good for the anti rainham.

556
00:49:50,930 --> 00:49:52,010
So it's even the right colours.

557
00:49:52,640 --> 00:50:01,160
And up here he's taking something which is nice and nice and wiggly, so that we can see that these bits with bigger curvature are going to disappear.

558
00:50:01,520 --> 00:50:04,070
This is almost flat and it stays almost flat.

559
00:50:04,460 --> 00:50:09,380
So what this stone is trying to do, and it's more like the shape of the zones that we see in other natural populations,

560
00:50:10,670 --> 00:50:15,840
is it's trying to get to be got to be a straight line. Okay.

561
00:50:18,260 --> 00:50:23,169
So we can expect that in natural populations. Approximately.

562
00:50:23,170 --> 00:50:27,610
At least if we look over large scales, things are going to evolve according to this curvature flow.

563
00:50:28,060 --> 00:50:32,780
It's probably the slowest example known of curvature flow, but it's rather cute.

564
00:50:32,800 --> 00:50:35,980
I think it's rather nice that we we can see how these things will move.

565
00:50:36,010 --> 00:50:42,150
We have not tested yet against a against data because this simply isn't going to be enough.

566
00:50:42,160 --> 00:50:45,630
It's going to move very, very slowly. So.

567
00:50:46,650 --> 00:50:53,610
We've assumed that those two populations that have come back together and then interfaced that the hybrids are not as fit as the purebreds.

568
00:50:53,610 --> 00:51:01,230
But we've assumed that the pure populations are equally fit. In fact, very often you can expect that the pure populations will not be equally fit.

569
00:51:01,980 --> 00:51:06,629
And if the populations are not equally fit. So I'm just showing off now, writing down random differential equations.

570
00:51:06,630 --> 00:51:14,040
Of course, the equation we write down before the island colony equation this year was equal to one, and that made this symmetric about a half.

571
00:51:14,340 --> 00:51:21,930
So that said, the frequency of different illegals in our population was symmetric about a half.

572
00:51:22,320 --> 00:51:28,380
And when it's bigger than a half, this time pushes me towards one, and then it's less than a half this term push, it pushes me towards zero.

573
00:51:28,860 --> 00:51:34,020
And so this gives me the competition between dispersion, dispersal and selection.

574
00:51:34,710 --> 00:51:38,340
If I is not equal to one, that's what happens when the populations are not equally fit.

575
00:51:38,640 --> 00:51:43,530
Then we get something called the Con Hillyard equation, which I can expand out as having a symmetric term.

576
00:51:43,860 --> 00:51:50,790
And this term here, which is no longer, no longer pushing me towards zero and one.

577
00:51:53,120 --> 00:51:58,280
And this we can expect to model that situation where things are not equally fit and in fact, we add a noise to them.

578
00:51:59,030 --> 00:52:05,749
And what happens in a situation is that the fighter type is going to spread in a travelling wave and it's happening much faster than curvature flow.

579
00:52:05,750 --> 00:52:11,750
It's going to it's going to spread across the, the range of the species as a travelling wave on a much faster timescale.

580
00:52:12,380 --> 00:52:13,970
Now why am I showing you this?

581
00:52:13,970 --> 00:52:22,030
Actually the two interesting because for any PDA people, it is interesting that if this time we're not here, then we get just pure selection.

582
00:52:22,040 --> 00:52:28,489
So for a haploid population, as we call it, so for populations where there's only one copy of each gene in each individual,

583
00:52:28,490 --> 00:52:30,890
and we're just saying that one type is fitter than the other type,

584
00:52:31,340 --> 00:52:35,899
this time wouldn't be here and this travelling wave would still exist, but it would spread out to the right,

585
00:52:35,900 --> 00:52:42,740
roughly the square root of S times, a minus one for these populations where we've also got the selection against hybridisation.

586
00:52:44,420 --> 00:52:47,989
The travelling wave will travel at a speed proportional to s times, a minus one,

587
00:52:47,990 --> 00:52:53,180
so a completely different speed, which is because we have a pushed wave instead of a pooled wave.

588
00:52:53,420 --> 00:52:56,720
So that was just proving. I've read some mathematics that was for Alan's benefit.

589
00:52:57,600 --> 00:53:00,680
So why do mathematicians like equations like this?

590
00:53:00,690 --> 00:53:06,150
We like equations like this and we like models like this because we recognise them as having come up elsewhere.

591
00:53:06,510 --> 00:53:14,170
So this we've come across for biological reasons. It's actually an equation which has been studied extensively, especially in physics.

592
00:53:14,590 --> 00:53:18,460
And we like that kind of universality where particular models arise in lots of different contexts.

593
00:53:18,790 --> 00:53:23,800
And the only difference here is that the form of the noise we take is really very different from the form of the noise that the physicists use.

594
00:53:24,640 --> 00:53:28,030
And the other thing that mathematicians like is they always like an excuse to mention coffee,

595
00:53:28,180 --> 00:53:31,690
an especially excuse to drink coffee, and even more so an excuse to spill coffee.

596
00:53:32,380 --> 00:53:39,940
And it turns out that this equation will model is travelling front and the fluctuations in that travelling front should

597
00:53:39,940 --> 00:53:45,580
be roughly the same as the fluctuations in the travelling front when you spill your coffee all over your exam scripts.

598
00:53:46,180 --> 00:53:49,210
So on that note, I think I'll stop. Thank you very much.