Transcript Document

The relative role of drift and selection in life-history
evolution: a case study from recently founded
populations of grayling
Mikko Koskinen, Thrond Haugen & Craig Primmer
Outline
• Background to ‘Darwinian’ vs ‘neutral’ evolutionary
theories
• Presentation of the framework of this study
• Results and conclusions
Natural Selection vs Random Drift?
• Transplantation experiments suggest that selection is an
efficient evolutionary force
- Anolis lizards were
introduced onto small islands
- Populations differentiated from
each other over 10-14 years according
to the recipient island’s vegetation
(Losos et al. 1997, Nature)
Natural Selection vs Random Drift?
• Comparisons of phenotypic differentiation and selectively
neutral differentiation (e.g. from non-coding DNA
microsatellites) suggest that selection is an efficient
evolutionary force
1
QS
- Mean quantitative genetic (QST)
differentiation generally exceeded
neutral marker gene (FST) differentiation
0.1
T
(Merilä & Crnokrak 2001, J. Evol. Biol.)
0.01
0.01
0.1
FST
1
Natural Selection vs Random Drift?
S. Wright
• In stark contrast with the Darwinian view, influential
theories (e.g. Wright 1931, Kimura 1995) suggest that drift is
the dominant evolutionary force in finite populations
• It is fair to say that the ‘neutral drift hypothesis’ has been
among the most controversial issues in evolutionary biology
in the last 50 years, and is empirically understudied
Outline
• Background to ‘Darwinian’ vs ‘neutral’ evolutionary
theories
• Presentation of the framework of this study
• Results and conclusions
Study system
European grayling,
Thymallus thymallus
Norway
The ‘plan’
• To measure quantitative genetic differences (QST) between the populations
using ‘common-garden’ experiments (six early life-history traits)
• Three temperatures, three populations with half-sib design
• Four unique females mated with each male (28 families per population)
2
 GB
QST  2
2
 GB  2 GW
(Spitze et al 1993, Genetics)
Variance components from mixed-model Anova
95% CI from non-parametric bootstrapping
• To measure neutral genetic differences (FST), i.e. the effect of drift, using 17
microsatellite DNA loci
The ‘plan’
• To investigate the demographic history of the populations using
microsatellites, and to use that for interpreting how the results relate to
the ‘Darwinian’ vs ‘neutral’ evolutionary theories
• To test the null-hypothesis of neutral evolution of the six traits using:
F1,
2
N e GB
 2 2
h  GW t
(Lande 1976, Evolution)
-Ne = effective population size (maximum-likelihood estimate from
microsatellite data)
-2GB = additive genetic variance between populations
-2GW = additive genetic variance within populations (among sire var comp)
-h2 = narrow-sense heritability in a given population and environment
-t = divergence time of populations
Outline
• Background to ‘Darwinian’ vs ‘neutral’ evolutionary
theories
• Presentation of the framework of this study
• Results and conclusions
Results - neutrality tests
• Neutral evolution was rejected for the majority of the trait
• Recall:
F1,
2
N e GB
 2 2
h  GW t
Trait
F
P
length at termination
384
<0.0001
yolk-sac volume
647
<0.0001
growth rate
111
<0.0001
incubation time
17.4
<0.0001
swim-up length
2.50
0.1235
hatching length
0.70
0.5472
Results - neutrality tests
• Extremely low Ne estimates, not compatible with sexual
reproduction, would have been required for drift to
dominate over selection
F1,  = 3.84 for P  0.05
Trait
F
P
Ne(sign)
length at termination
384
<0.0001
0.25
yolk-sac volume
647
<0.0001
0.14
growth rate
111
<0.0001
0.84
incubation time
17.4
<0.0001
5.33
swim-up length
2.50
ns
37.1
hatching length
0.70
ns
218
Results - QST vs FST
• Population differences based on quantitative traits (QST)
often strikingly exceeded the analogous measures based on
microsatellites (FST)
Les vs Ht
FST
length at termination
yolk-sac volume
growth rate
incubation time
swim-up length
hatching length
0.0
0.5
1.0
Results - demographic history
• Effective sizes of the populations were small
• Microsatellite diversity within populations was low
• The populations have historically experienced severe
‘bottlenecks’
Population
n
Ne (95% CI)
A (SD)
N0 / N1 (95% CI)
Les
52
88.9
1.9 (1.1)
0.003 (0.0003-0.03)
Ht
48
24.2 (12.1-42.2)
1.5 (0.9)
0.006 (0.0003-0.05)
ØM
49
85.0 (36.0-170.5)
1.7 (0.9)
0.0006 (0.00008-0.01)
Aur
28
55.4 (24.8-110.0)
1.6 (0.9)
0.001 (0.0003-0.005)
Show examples of Ne sampling distributions and likewise for N0/N1
Conclusions
• The evolution of the phenotypic differences between the
populations was dominantly due to natural selection
-neutrality tests [F-test and Ne(sign) estimates]
-QST vs FST comparisons
• Provide Fst/Qst+Fst
Conclusions
• However, also drift had a notable effect (FST = 0.05-0.21)
• The dominating effect of selection is interesting in the light
of the demographic history of the populations.
• According to the influential ‘neutral theory’, the low Nes
and bottlenecks should have emphasized the effect of drift
Acknowledgements
• Thanks to Juha Merilä, Mark Beaumont, Asbjørn
Vøllestad, Peter Crnokrak, Andrew Hendry, Martin
Lascoux and Nick Smith for helpful comments!