Explaining the evolution of sex and recombination

Explaining the evolution of
sex and recombination
Peter Keightley
Institute of Evolutionary Biology
University of Edinburgh
Sexual reproduction is ubiquitous in
eukaryotes
Syngamy
Meiosis with
recombination
Sex and recombination have potential
disadvantages
●
Meiotic recombination breaks apart
combinations of genes that have been built up
by selection.
Sex and recombination have potential
disadvantages
●
●
Meiotic recombination breaks apart
combinations of genes that have been built up
by selection.
Sexual reproduction requires
finding a mate, which carries risks.
Sex and recombination have potential
disadvantages
●
●
●
Meiotic recombination breaks apart
combinations of genes that have been built up
by selection.
Sexual reproduction requires
finding a mate, which carries risks.
Sex can carry a two-fold cost
if population growth is limited by
the number of females.
●
Why are sex and
recombination almost
ubiquitous in nature?
●
●
Why are sex and
recombination almost
ubiquitous in nature?
Why don’t asexuals
take over sexual
populations?
Is there a general explanation for the
evolution of recombination?
Collaborator: Sally Otto
University of British Columbia
Association between beneficial (+) and
deleterious alleles (-)
+
-
-
+
-
+
+
-
Four
chromosomes
in a population
Association between beneficial (+) and
deleterious alleles (-)
●
+
-
-
+
-
+
+
-
Alleles at the loci are negatively associated.
●
The coupling types ++ and -- are absent.
Association between beneficial (+) and
deleterious alleles (-)
●
-
-
+
-
+
+
-
Alleles at the loci are negatively associated.
●
●
+
The coupling types ++ and -- are absent.
The effectiveness of selection acting on variation at
the loci is weakened.
Mutations that increase recombination are favoured if
there are negative allelic associations
+
-
m
-
+
m
-
+
m
+
-
m
Mutations that increase recombination are favoured if
there are negative allelic associations
+
-
M
-
+
m
-
+
m
+
-
m
Mutations that increase recombination are favoured if
there are negative allelic associations
●
+
-
M
-
+
m
-
+
m
+
-
m
+
+
M
A mutation (M) at a linked locus that increases
recombination is favoured because it hitchhikes on
chromosomes carrying multiple beneficial alleles,
generated by recombination.
Felsenstein & Yokoyama 1976
Non-random associations between linked loci arise in
two ways
1. Directional selection.
+
+
-
+
+
-
-
-
size is very large.
Rapidly fixed
Rapidly eliminated
Non-random associations between linked loci arise in
two ways
1. Directional selection.
size is very large.
-
+
+
-
Non-random associations between linked loci arise in
two ways
1. Directional selection.
size is very large.
-
+
+
-
BUT non-random associations are
expected to be absent in large
populations if there is some
recombination.
Disequilibrium
Pop. size
Non-random associations between linked loci arise in
two ways
2. Synergistic epistatic interactions between loci.
+
+
-
+
+
-
-
-
Unfit combination that
is rapidly eliminated
Non-random associations between linked loci arise in
two ways
2. Synergistic epistatic interactions between loci.
+
+
-
+
+
-
BUT there is little empirical evidence for this kind of epistasis.
Is recombination favoured in the presence
of selection at multiple linked loci?
Simulating the fate of a recombination modifier in the
presence of selection and mutation at multiple linked loci
Population of N haploid
individuals at equilibrium
between linked
deleterious mutations
and selection.
Simulating the fate of a recombination modifier in the
presence of selection and mutation at multiple linked loci
M
Introduce a mutation
that increases
recombination.
Frequency = 1/N
Possible fate of a recombination modifier mutation
Modifier
lost
M
Possible fate of a recombination modifier mutation
Modifier
lost
M
Modifier fixed
MM MM
M M
M M M M M Happens at
MMM M
M a frequency of
M M
M MM M 1/N by chance
MM
●
Modifiers mutations that increase
recombination were always observed to be
favoured
Recombination is strongly favoured in
nonrecombining systems
Relative fixation rate of modifier
100
Recombination
modifier
increasing map
length by 0.1M
10
1
0
0.25
0.5
0.75
1
Initial recombination rate (map length, L)
The advantage of recombination increases with
population size
Relative fixation rate of modifier
100
Map length
L=0
L = 0.1
L=1
10
1
0.1
0
5000 10000 15000 20000 25000
Population size, N
The advantage of recombination increases with
population size
Relative fixation rate of modifier
100
Map length
L=0
L = 0.1
L=1
10
1
0.1
0
5000 10000 15000 20000 25000
Population size, N
●
This property “emerges” because of the multi-locus
nature of the model.
●
In systems with small numbers of loci, recombination is
favoured only in small populations.
Why does the advantage of a recombination
modifier increase with population size?
●
Selection generates multi-locus associations.
Why does the advantage of a recombination
modifier increase with population size?
●
●
Selection generates multi-locus associations.
Large populations maintain more polymorphic loci, and more
negative associations, increasing the hitchhiking effect favouring
recombination modifier mutations.
Why does the advantage of a recombination
modifier increase with population size?
●
●
●
Selection generates multi-locus associations.
Large populations maintain more polymorphic loci, and more
negative associations, increasing the hitchhiking effect favouring
recombination modifier mutations.
The infinite population prediction that negative associations
should be absent does not apply when there are multiple loci
under selection.
Summary
●
Simple simulations show that interference between
selected mutations favours mutations that increase
recombination.
Summary
●
●
Simple simulations show that interference between
selected mutations favours mutations that increase
recombination.
Theory developed for two- or three-locus systems proved
to be misleading.
Summary
●
●
●
Simple simulations show that interference between
selected mutations favours mutations that increase
recombination.
Theory developed for two- or three-locus systems proved
to be misleading.
The mechanism described is likely to be the general
explanation for the evolution of recombination.
Mutations affecting fitness are universal and natural
populations tend to be very large.