matthew m. osmond & sarah p. otto · 2020. 11. 5. · crossing tness-valleys without mendel matthew...
TRANSCRIPT
-
Crossing fitness-valleys without Mendel
Matthew M. Osmond & Sarah P. Otto
Biodiversity Research Centre & Department of ZoologyUniversity of British Columbia, Vancouver, Canada
-
Mendel’s law of segregation
Aa
A a
Parent
Gametes
1/2 1/2 Probability
-
Mendel’s law of segregation
Aa
A a
Parent
Gametes
1/2 1/2 Probability
-
Mendel’s law of segregation
Aa
A a
Parent
Gametes
1/2 1/2 Probability
-
Mendel’s law of segregation
Aa
A a
Parent
Gametes
1/2 1/2 Probability
-
Non-mendelian transmission
//drive// significant portion of any eukaryote genome is com-posed of selfish elements that gain a transmission advantage(Hurst & Werren 2001)
//uniparental inheritance// mitochondria and chloroplastgenes often inherited from parent of one particular sex(Birky Jr. 2001)
//non-genetic inheritance// increasing awareness of epige-netics, parental effects, and ecological and cultural inheritance(Danchin et al. 2011)
-
Non-mendelian transmission
//drive// significant portion of any eukaryote genome is com-posed of selfish elements that gain a transmission advantage(Hurst & Werren 2001)
//uniparental inheritance// mitochondria and chloroplastgenes often inherited from parent of one particular sex(Birky Jr. 2001)
//non-genetic inheritance// increasing awareness of epige-netics, parental effects, and ecological and cultural inheritance(Danchin et al. 2011)
-
Drive e.g.: spore killers in fungi
Raju 1980, Saupe 2012
-
Non-mendelian transmission
//drive// significant portion of any eukaryote genome is com-posed of selfish elements that gain a transmission advantage(Hurst & Werren 2001)
//uniparental inheritance// mitochondria and chloroplastgenes often inherited from parent of one particular sex(Birky Jr. 2001)
//non-genetic inheritance// increasing awareness of epige-netics, parental effects, and ecological and cultural inheritance(Danchin et al. 2011)
-
Uniparental inheritance e.g.: human mitochondria
Mitochondria is alwaysinherited from the mother
http://evolution.berkeley.edu
-
Non-mendelian transmission
//drive// significant portion of any eukaryote genome is com-posed of selfish elements that gain a transmission advantage(Hurst & Werren 2001)
//uniparental inheritance// mitochondria and chloroplastgenes often inherited from parent of one particular sex(Birky Jr. 2001)
//non-genetic inheritance// increasing awareness of epige-netics, parental effects, and ecological and cultural inheritance(Danchin et al. 2011)
-
Non-genetic inheritance e.g.: culture
Cavalli-Sforza & Feldman 1981
-
Non-genetic inheritance e.g.: culture
Cavalli-Sforza & Feldman 1981
-
Non-mendelian transmission is everywhere
//drive// significant portion of any eukaryote genome is com-posed of selfish elements that gain a transmission advantage(Hurst & Werren 2001)
//uniparental inheritance// mitochondria and chloroplastgenes often inherited from parent of one particular sex(Birky Jr. 2001)
//non-genetic inheritance// increasing awareness of epige-netics, parental effects, and ecological and cultural inheritance(Danchin et al. 2011)
-
Epistasis is also everywhere
Non-linear interactionsbetween traits
Phillips 2008Phillips 2008
-
Epistasis is also everywhere
Non-linear interactionsbetween traits
Phillips 2008Phillips 2008
-
Epistasis is also everywhere
Non-linear interactionsbetween traits
Phillips 2008
Phillips 2008
-
Epistasis is also everywhere
Non-linear interactionsbetween traits
Phillips 2008
Phillips 2008
-
Epistasis causes peaks and valleys in fitness
Wright 1932
-
Crossing valleys: the shifting balance theory
Wright 1932
-
The problem
1. Non-Mendelian inheritance is everywhere
2. Epistasis is pervasive, causes fitness valleys
3. Current valley-crossing theory limited to Mendelian traits
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
The problem
1. Non-Mendelian inheritance is everywhere
2. Epistasis is pervasive, causes fitness valleys
3. Current valley-crossing theory limited to Mendelian traits
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
The problem
1. Non-Mendelian inheritance is everywhere
2. Epistasis is pervasive, causes fitness valleys
3. Current valley-crossing theory limited to Mendelian traits
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
The problem
1. Non-Mendelian inheritance is everywhere
2. Epistasis is pervasive, causes fitness valleys
3. Current valley-crossing theory limited to Mendelian traits
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
The problem
1. Non-Mendelian inheritance is everywhere
2. Epistasis is pervasive, causes fitness valleys
3. Current valley-crossing theory limited to Mendelian traits
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
Our simple fitness valley
A1B1
A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
-
Our simple fitness valley
A1B1 A2B1 A1B2
A2B2
W21
W12
W11
W22
Genotype
Fit
ness
-
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
-
Generalized inheritance
A1B1 X A2B2
A1B1 A2B1 A1B2 A2B2
FatherMother
Offspring
12(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)1
2(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)
b2211(11) b2211(21) b
2211(12) b
2211(22)
-
Generalized inheritance
A1B1 X A2B2
A1B1 A2B1 A1B2 A2B2
FatherMother
Offspring
12(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)1
2(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)
b2211(11) b2211(21) b
2211(12) b
2211(22)
-
Generalized inheritance
A1B1 X A2B2
A1B1 A2B1 A1B2 A2B2
FatherMother
Offspring
12(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)
12(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)
b2211(11) b2211(21) b
2211(12) b
2211(22)
-
Generalized inheritance
A1B1 X A2B2
A1B1 A2B1 A1B2 A2B2
FatherMother
Offspring
12(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)
12(1 − r − µ) 1
2(r + µ) 1
2(r + µ) 1
2(1 − r − µ)
b2211(11) b2211(21) b
2211(12) b
2211(22)
-
Methods
Markov Diffusion Results
-
Drive
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Val
ley-
cros
sing
tim
e
Probability of recombination
Drive has large
st effect
when crossing s
low
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Drive
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Val
ley-
cros
sing
tim
e
Probability of recombination
Drive has large
st effect
when crossing s
low
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Drive
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Val
ley-
cros
sing
tim
e
Probability of recombination
Drive has large
st effect
when crossing s
low
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Drive
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Val
ley-
cros
sing
tim
e
Probability of recombination
Drive has large
st effect
when crossing s
low
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Drive
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Drive favours residents
No drive
Drive favours mutants
10-6
10-5
10-4
0.001 0.01 0.1r0
10 000
20 000
30 000
40 000
T
Val
ley-
cros
sing
tim
e
Probability of recombination
Drive has large
st effect
when crossing s
low
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fit
ness
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Uniparental inheritance
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Increasing mut
ation
speedscrossing
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Mutational asym
metry
hinderscrossing
Val
ley-
cros
sing
tim
e
Ratio of mutation probability (µB/µA)
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Uniparental inheritance
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Increasing mut
ation
speedscrossing
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Mutational asym
metry
hinderscrossing
Val
ley-
cros
sing
tim
e
Ratio of mutation probability (µB/µA)
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Uniparental inheritance
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Increasing mut
ation
speedscrossing
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Mutational asym
metry
hinderscrossing
Val
ley-
cros
sing
tim
e
Ratio of mutation probability (µB/µA)
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Uniparental inheritance
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Increasing mut
ation
speedscrossing
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Mutational asym
metry
hinderscrossing
Val
ley-
cros
sing
tim
e
Ratio of mutation probability (µB/µA)
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Uniparental inheritance
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Increasing mut
ation
speedscrossing
0.01 0.1 1 10 100cΜ
1000
104
105
106
T
Mutational asym
metry
hinderscrossing
Val
ley-
cros
sing
tim
e
Ratio of mutation probability (µB/µA)
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Non-genetic inheritance
Small double mutant advantage Hk22=1100L
Large double mutant advantage Hk22=110L
0.0 0.1 0.2 0.3 0.4 0.5 0.6q1
10
100
1000
104
105T
Small double mutant advantage Hk22=1100L
Large double mutant advantage Hk22=110L
0.0 0.1 0.2 0.3 0.4 0.5 0.6q1
10
100
1000
104
105T
Val
ley-
cros
sing
tim
e
Transmission probability of single mutants
Transmission a
dvantage
of newcombin
ation has
large relative e
ffect
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Non-genetic inheritance
Small double mutant advantage Hk22=1100L
Large double mutant advantage Hk22=110L
0.0 0.1 0.2 0.3 0.4 0.5 0.6q1
10
100
1000
104
105T
Small double mutant advantage Hk22=1100L
Large double mutant advantage Hk22=110L
0.0 0.1 0.2 0.3 0.4 0.5 0.6q1
10
100
1000
104
105T
Val
ley-
cros
sing
tim
e
Transmission probability of single mutants
Transmission a
dvantage
of newcombin
ation has
large relative e
ffect
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Non-genetic inheritance
Small double mutant advantage Hk22=1100L
Large double mutant advantage Hk22=110L
0.0 0.1 0.2 0.3 0.4 0.5 0.6q1
10
100
1000
104
105T
Small double mutant advantage Hk22=1100L
Large double mutant advantage Hk22=110L
0.0 0.1 0.2 0.3 0.4 0.5 0.6q1
10
100
1000
104
105T
Val
ley-
cros
sing
tim
e
Transmission probability of single mutants
Transmission a
dvantage
of newcombin
ation has
large relative e
ffect
Our simple fitness valley
A1B1 A2B1 A1B2 A2B2
W21
W12
W11
W22
Genotype
Fitne
ss
-
Conclusions
1. DriveI helps or hindersI largest effect when valley crossing otherwise slow
2. Uniparental inhertianceI implies mutational asymmetry with biparental traitI increased mutation helps valley crossingI mutational asymmetries hinder valley crossing
3. Non-genetic inheritanceI transmission advantage of new combination has large
relative effect on valley crossing times
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
Conclusions
1. DriveI helps or hindersI largest effect when valley crossing otherwise slow
2. Uniparental inhertianceI implies mutational asymmetry with biparental traitI increased mutation helps valley crossingI mutational asymmetries hinder valley crossing
3. Non-genetic inheritanceI transmission advantage of new combination has large
relative effect on valley crossing times
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
Conclusions
1. DriveI helps or hindersI largest effect when valley crossing otherwise slow
2. Uniparental inhertianceI implies mutational asymmetry with biparental traitI increased mutation helps valley crossingI mutational asymmetries hinder valley crossing
3. Non-genetic inheritanceI transmission advantage of new combination has large
relative effect on valley crossing times
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
Conclusions
1. DriveI helps or hindersI largest effect when valley crossing otherwise slow
2. Uniparental inhertianceI implies mutational asymmetry with biparental traitI increased mutation helps valley crossingI mutational asymmetries hinder valley crossing
3. Non-genetic inheritanceI transmission advantage of new combination has large
relative effect on valley crossing times
Q. How does non-Mendelian inheritance affectfitness-valley crossing?
-
Thank-you!
I The Otto & Whitlock labs
I The Doebeli & Hauert labs
I The Kisdi & Geritz labs
I Mark Kirkpatrick
BackgroundProblemModelResultsConclusionsAcknowledgements