Transcript Adaptive variation

Adaptive variation
A feature of an organism that has been favoured by
natural selection because of that feature's positive effect
on relative fitness
Identifying local adaptation
Common garden experiments
Clines
Qst (phenotypic differentiation) versus Fst (genetic
differentiation at neutral molecular markers)
The definition of local adaptation (Kawecki
& Ebert
Annual
Reviews2004).
Common garden experiments
Potentilla glandulosa
Clausen, Keck, & Hiesey
Common gardens
Phenotypic plasticity
Annual Reviews
Transfer response functions for fitness and its components in Pinus sylvestris for
a central population from latitude 60◦N and a northern population from latitude 66◦N.
Annual Reviews
Clinal variation in traits related to timing of growth i

Vw= average within population genetic variation

Vb= average between population genetic variation

Qst=Vb/(Vw+2Vb)

Note these are genetic variances, not phenotypic
variances

Need estimates of heritability within populations

Clonal Daphnia used by Spitze
Annual Reviews
FST and QST values of twelve tree species
Not all traits are adaptations
Neutral processes (e.g. drift)
Exaptations - a trait may have evolved previously for
another purpose (Gould)
Pleiotropy - selection on another trait which is
controlled by the same genes
Phenotypic plasticity
Historical contingency (multiple adaptive peaks)
Effects of climate change on plant
populations
Climate change may occur more quickly than migration
The degree of phenotypic plasticity may be less than is required
to deal with the climatic variability associated with climate
change
Can plants adapt to climate change?
Effects of climate change on plant
populations: adaptation
Habitat fragmentation
-Ne reduced (drift increases, efficiency of selection reduced)
-reduced gene flow (m<1)
-erodes genetic variation, increased inbreeding (inbreeding
depression) -> reduced population fitness
Strong selection pressures from multiple sources may exhaust
genetic variation -> population can’t stay at fitness optima
Genetic correlations among traits can impede the response to
selection
Species with long generation times will respond slowest
Genetic variation and extinction risk


A small population is prone to positive
feedback loops in inbreeding and
genetic drift that draws the population
down an extinction vortex toward
smaller and smaller population size
until extinction (mutational meltdown)
Thus the rate of adaptation may be
outstripped by climate change for many
species->extinction
Outlier FST as evidence for adaptive variation
Locations of the 6 sampled populations
Success of SNP assays
Summary statistics by
population
Analysis for adaptive
differentiation

The program “newfst” (Beaumont & Balding 2004) was used to
identify genes subject to selection

This program relies on a Bayesian model to generate FST values
through a Markov Chain Monte Carlo (MCMC) algorithm

It can disentangle the locus effect (αi), the population effect (βj), and
the interaction between the locus and the population effects (γij).

A large positive αi indicates the presence of a positive selection on
the studied gene, while a large positive γij indicates locus–population
interaction, thus a potentially advantageous mutation that would be
locally adapted to a particular population

Loci with high positive γij values (above 0.10) possibly reflect true
adaptive differentiation
Obtain estimates of F for locus i, population j.
Fit the following linear model:
 is locus effect (averaged over populations)
 is population effect (averaged over loci)
 is locus x population effect (adaptation in specific populations)
It is possible to identify the majority of loci under adaptive selection; in
simulations, good discrimination for adaptively selected loci when s > 5m.
(s = selection coefficient, m = migration rate among populations)
Back to Namroud et al.
Conclusions of Namroud et al.







First genome-wide SNP scan of genes in a nonmodel species
First to be conducted in conifer populations for which significant
genetic differentiation in quantitative traits has been demonstrated
from common garden studies
Average among-population FST was very low (0.006)
No strong local adaptation (no positive γij at the 95% or the 99%
confidence levels), but 49 SNPs showed a “trend” towards local
adaptation (γij value > 0.10), despite low FST .
“Ascertainment bias”: Only SNPs of higher frequency were assayed,
yet low frequency SNPs might contribute most to local adaptation
Clear definition of phyiological roles of these SNPs is a long way
from being determined (need association, functional studies)
“Next generation” sequencing methods will make sequencing and
genotyping much less expensive
Genecology and Adaptation
of Douglas-Fir to Climate
Change
Brad St.Clair1, Ken Vance-Borland2 and Nancy Mandel1
1USDA Forest Service, Pacific Northwest Research Station
2Oregon State University
Corvallis, Oregon
Objectives of this study
To explore geographic genetic structure and
the relationship between genetic variation and
climate
To evaluate the effects of changing climates on
adaptation of current populations
To consider the locations of populations that
might be expected to be best adapted to future
climates
Genecology
Definition: the study of intra-specific genetic variation
of plants in relation to environments (Turesson 1923)
Consistent correlations between genotypes and
environments suggest natural selection and
adaptation of populations to their environments
(Endler 1986)
Methods for exploring genecology and geographic
structure – common garden studies
– Classical provenance tests
– Campbell approach
intensive sampling scheme
particularly advantageous in the highly heterogeneous
environments in mountains
Objective 1: Geographic structure and relationship
between genetic variation and climate
Douglas-fir common garden study
Raised beds
Distribution of parent
trees and elevation
Analysis
Canonical correlation analysis
– Determines pairs of linear combinations from two
sets of original variables such that the correlations
between canonical variables are maximized
– Trait variables
emergence, growth, bud phenology, and partitioning
– Climate variables
modeled by PRISM
annual and monthly precipitation, minimum and maximum
temperatures, seasonal ratios
Use GIS to display results
Results from CCA
Component
Canonical
Correlation
Canonical
R-squared
Proportion of
trait variance
explained by
CV for traits
Proportion of
trait variance
explained by
CV for climate
1
0.86
0.73
0.39
0.29
2
0.59
0.35
0.11
0.04
3
0.34
0.11
0.04
0.005
First component accounted for much of the variation.
First component may be called vigor – correlated with large size
(r=0.65), late bud-set (r=0.94), high shoot:root ratio (r=0.60),
and fast emergence rate (r=0.71).
Results from CCA
First CV for Traits correlated with:
Dec min temperature
0.79
Jan min temperature
0.73
Feb max temperature
0.73
Mar min temperature
0.77
Aug min temperature
0.42
Aug precipitation
0.30
Model:
trait1=-0.08+0.38*decmin –0.25*janmin+0.09*febmax
+0.13*marmin-0.12*augmin+0.02*augpre
Geographic genetic variation in
first canonical variable for traits
CV 1 for Traits
Dec Minimum Temperature
Objective 2: Effects of changing climates on
adaptation of current populations
Methods
1. Develop model of the relationship between
genetic variation and environment using climate
variables.
2. Given model, determine set of genotypes that
may be expected to be best adapted to future
climate.
3. Given climate change, determine degree of
maladaptation of current population to changed
climate (determined by the mismatch between
current population and best adapted population).
Climate change predictions
Two models:
– Canadian Center for Climate Modeling and Analysis
– Hadley Center for Climate Prediction and Research
We assumed no geographic variation in
climate change
Climate change predictions
Expected Values for Climate Change (ºC)
Jan
Min
Temp
Feb
Max
Temp
Mar Min
Temp
Aug
Min
Temp
Aug
Precip
(ratio)
C 2030 2.5
2.5
1.8
2.0
1.0
0.9
H 2030 2.3
2.3
1.7
2.1
1.8
1.0
C 2090 6.0
6.0
5.8
5.5
4.4
1.0
H 2090 5.5
5.5
4.0
5.2
4.7
0.9
Model/Year
Dec Min
Temp
Geographic genetic variation that may
be expected to be best adapted to
present and future climates
Present
2030
2095
Summary of Objective 2: Effects of
changing climates on adaptation of
current populations
40% risk of maladaptation within acceptable
limits of seed transfer (Campbell, Sorensen).
71-84% risk is somewhat high.
Enough genetic variation present to allow
evolution through natural selection or migration.
Maladaptation does not necessarily mean
mortality. Trees may actually grow better, but
below the optimum possible with the best
adapted populations.
Objective 3. To consider the locations of
populations that might be expected to be
best adapted to future climates
Focal Point Seed Zones
present
2030
2095
How far down in elevation do we go to find populations
adapted to future climates?
Year
2095
3
Year
2030
Year
2000
2
CV Trait 1
1
0
-1
-2
-3
-4
-5
0
200 400 600 800 1000 1200 1400 1600 1800 2000
Elevation
r = -0.69
Conclusions
Douglas-fir has considerable geographic genetic
structure in vigor, most strongly associated with winter
minimum temperatures.
Climate change results in some risk of maladaptation,
but current populations appear to have enough
genetic variation that they may be expected to evolve
to a new optimum through natural selection or
migration.
Populations that may be expected to be best adapted
to future climates will come from much lower
elevations, and, perhaps, further south.
Forest managers should consider mixing seed from
local populations with populations that may be
expected to be adapted to future climates.