Selection on Quantitative Traits

1
Selection on Quantitative Traits
2
Q1. Todays topic is.

1. Sexual selection
2. Heritability
3. Linkage Equilibrium
4. Fossil Record
5. Hardy Weinberg exceptions

3
Q2. This graph represents.

• Heritability
• Selection differential
• Selection gradient
• Directional selection
• Stabilizing selection

4
Q3.Candace Galen worked with which group of
organisms?

1. Flowers
2. Fruit flies
3. Beetles
4. Guppies
5. Newts

5
Quantitative traits are those that show
continuous variation

• height
• skin color
• plant flower size

Qualitative traits are all or none

• attached earlobes
• widows peak
• six fingers
• cystic fibrosis

6
Quantitative genetics

• provides tools to analyze genetics and evolution
  of continuously variable traits
• Provides tools for
• 1. measuring heritable variation
• 2. measuring survival and reproductive success
• 3. predicting response to selection

7
Measuring Heritable Variation Text section 9.3

• When assessing heritability we need to make
  comparisons among individuals. Cannot assess a
  continuous traits heritability within one
  individual
• Need to differentiate whether the variability we
  see is due to environmental or genetic
  differences
• Heritability The fraction of the total
  variation which is due to variation in genes

8
Types of Variation

• Phenotypic variation (VP) is the total variation
  in a trait (VE VG)
• Environmental variation. (VE) is the variation
  among individuals that is due to their
  environment
• Genetic variation (VG) is the variation among
  individuals that is due to their genes

9
Total genetic variation (VG) actually has two
components

• Additive Genetic Variation (VA) Variation among
  individuals due to additive effects of genes
• Dominance Genetic Variation (VD) Variation
  among individuals due to gene interactions such
  as dominance
• VG VA VD

10
Measuring heritability
VG VG VE
VG VP

• heritability
• Heritability is always between 0 and 1
• If the variability is due to genes then it makes
  sense to evaluate the resemblance of offspring to
  their parents

11
Heritability can be looked at from two
perspectives

• Broad sense heritability VG / VP
• Narrow sense heritability VA / VP
• We will deal only with narrow sense heritability
  h2
• Use of narrow sense heritability allows us to
  predict how a population will respond to
  selection

12
Calculating h2- Testing the relationship between
parents and offspring trait values

• Plot midpoint value for the 2 parents on x axis
  and mid-offspring value on y axis and draw a best
  fit line.
• This slope which is calculated by least squares
  linear regression is a measure of heritability
  called narrow-sense heritability or h2
• h2  is an estimate of the fraction of the
  variation among the parents that is due to
  variation in the parents genes
• Looking at a hypothetical population

13
If slope is near zero there is no resemblance
Figure 9.13a Pg 334
Evidence that the variation among parents is due
to the environment.
14
If this slope is near 1 then there is strong
resemblance
Evidence the variation among parents is due to
genes
15
Need to make sure that the environment is not
causing some of the variation because
environment runs in families too.

• Any study of heritability needs to account for
  possible environmental causes of similarity
  between parent and offspring.
• Take young offspring and assign them randomly to
  parents to be raised
• In plants, randomly plant seeds in a given field
• Example in text Song Sparrows studied by James
  Smith and Andre Dhondt.

16
Showed song sparrow chicks (eggs or hatchlings)
raised by foster parents resembled their
biological parents strongly and their foster
parents not at all
Figure 9.14 p. 335
17
Measuring differences in Survival and
Reproductive Success

• Done by measuring the strength of selection by
  looking at the differences in reproductive
  success.
• Basically we measure who survives, who doesnt,
  and then quantify the difference
• Example breeding mice with longer tails

18
Artificial selection for long tail length

• DiMasso and colleagues bred mice in order to
  select for longer tails
• Each generation they picked the 1/3 of the mice
  who had the longest tails and allowed them to
  interbreed
• Did this for 18 generations
• Calculated the strength of selection

19
Two measures of the strength of selection

• Selection differential (S) difference between
  mean tail length of breeders (those that survive
  long enough to breed) and the mean tail length of
  the entire population.
• Selection gradient slope of a best fit line on
  a scatter plot of relative fitness as a function
  of tail length

20
Selection differential (S)
Figure 9.17 p. 339
Only the 1/3 of mice with the longest tails
allowed to breed (survive)
Average tail length of the breeders only minus
the average tail length of the entire population
entire population
breeders (survivors)
21
To calculate the selection gradient

1. Assign absolute fitness fitness equals survival
   to reproductive age. Long tailed had a fitness
   of 1, short tailed a fitness of 0
2. Convert absolute fitness to relative fitness.
   Figure the mean fitness of the population. Then
   divide the absolute fitness by the mean fitness .
   (Mean fitness .67(0) .33(1) .33). So
   relative fitness of breeders 1/.33 3.0 and
   relative fitness of non-breeders 0/.33 0.
3. Make a scatterplot of relative fitness as a
   function of tail length. Calculate the slope
   using best fit. The slope is the selection
   gradient

22
Selection gradient
Figure 9.17 p. 339
1. Calculate relative fitness for each mouse,
then plot relative fitness of each as a function
of tail length
2. the slope of the best fit line is the
selection gradient
23
Selection gradient is more useful because it can
be used to calculate any measure of fitness not
just survival

• Selection differential can be calculated from
  selection gradient
• Divide the selection gradient by the variance.
  Explained in box 9.3 p. 340.

24
Predicting Evolutionary response

• Once we know the heritability and the strength of
  selection we can predict response to selection
• R h2 S
• R predicted response
• h2 heritability
• S selection differential

25
Review of what we can do with the tools of
quantitative genetics.

• We can estimate how much variation in a trait is
  due to the variation in a gene (heritability)
• Quantify the strength of selection that results
  from differences in survival or reproduction.
  (selection differential)
• Predict how much a population will change from
  one generation to the next. (predicted response
  to selection)

26
Alpine Skypilot (Polemonium viscosum) studies
27
Alpine skypilots and Bumblebees

• Candace Galen (1966) studied the effect of
  selection pressure by bumblebees on flower
  diameter
• Worked with alpine skypilots from two elevations,
  timberline and tundra
• Tundra flowers are larger and are pollinated
  exclusively by bumblebees
• Timberline flowers are pollinated by a mixture of
  insects and are smaller

28
Galen wanted to determine two things

1. Is selection by the bumblebees in the tundra
   responsible for the larger flower size?
2. How long would it take for selection pressure to
   increase flower size by 15

29
Determining the response to selection

• Determine heritability
• measured flower diameters
• collected seeds germinated them and
  transplanted seedlings to random locations in the
  same habitat as the parents
• seven years later measured the flowers from
  the 58 plants which had matured enough to flower
• plotted offspring flower diameter as a
  function of maternal (seed bearing parent) flower
  diameter

30
Analysis of results

• results provided a best fit number of 0.5 for
  heritability. Actual calculations give h2 of 1.0
  (because multiple offspring with only one parent
  female).
• Scatter (fig 9.20) necessitated a statistical
  analysis which showed she could only be certain
  that at least 20 of the phenotypic variation was
  due to additive genetic variation. (h2 VA /
  VP)

Therefore h2 lies somewhere between 0.2 and 1.0
31
2. Determine selection differential

• caged some about-to-flower Skypilots with
  bumblebees
• measured flower size when Skypilots bloomed and
  later collected their seeds
• planted seedlings back out in the original
  parental habitat
• Six years later she counted the number of
  surviving offspring produced by each of the
  parent plants She used the number of surviving 6
  year old offspring as her measure of fitness
• Plotted relative fitness ( of surviving 6 year
  old offspring / total number planted) as a
  function of maternal flower size.

32
pg 343 Fig 9.21
33
Slope of best fit line is selection gradient

• Calculated the selection differential (S) ( by
  dividing selection gradient by variance in flower
  size)
• Her S value told her that, on average, the
  flowers visited by bumblebees were 5 larger than
  the average flower size.
• Control experiments from random hand pollination
  and by a mixture of pollinators other than
  bumblebees, showed no relationship between flower
  size and fitness

34
Fig 9.22 pg 343
35
3. Galen predicted a response

• using the low end h2 of .2 and an S of .05
• R h2S .2 (.05) .01
• using a high end for h2 of 1.0 and S .05
• R h2 S 1(.05) .05
• Means that a single generation of selection
  should produce an increase in the size of the
  average flower by from 1 to 5.

36

• Observations of a population of timberline
  flowers pollinated exclusively by bumblebees
  showed that on average flowers that were produced
  by bumblebee pollination were 9 larger than
  those pollinated randomly by hand.
• Galens prediction that response was rapid was
  verified

37
End for today
38
Modes of Selection on continuous traits

• How does each impact the distribution, fitness
  and survival of the possible phenotypes?

39
Directional Selection

• Fitness of a phenotype increase or decreases with
  the value of a trait.

Examples of this type of selection are The Alpine
Skypilot and the Finch beaks in times of drought.
One extreme phenotypic expression of the trait
increases in fitness and the other extreme
decreases.
Slightly reduces the variation in a population
40
Stabilizing Selection

• Those individuals with intermediate values are
  favored at the expense of both extremes.
• The average value of a trait remains the same but
  the variation is reduced
• The tails of the distribution are cut off.

41
Example in gall flies - Weis and Abrahamson 1986
Figure 9.26 p. 348

• A fly lays eggs in Goldenrod bud.
• Plant produces a gall in response to the fly
  larva
• Wasps lay eggs in galls that eat fly larva
• Birds also eat galls.
• Pressure from wasps selects for larger galls and
• Pressure from birds selects for smaller galls
• The result is selection for mid sized galls.

42
Disruptive Selection

• Selects for individuals with extreme values for a
  trait
• Does not change AVERAGE value but INCREASES
  phenotypic variance
• Result far fewer individuals at the middle of the
  continuum for the trait

43
Example of the black-bellied seed cracker

• Breeding Populations have birds with EITHER large
  OR small beaks
• Juveniles show the full spectrum of beak size
• But only the large OR small beaked birds survive
  to reproduce.

Fiogure 9.27 p. 349
44
Summary

• Unlike our example of the moths and other ONE
  gene traits.
• We are talking here about quantitative traits
  determined by multiple genes
• As phenotypic variation decreases so should
  genetic variation
• However in most populations substantial genetic
  variation continues to be exhibited.
• A satisfactory explanation for this unexpected
  outcome is under debate and no acceptable
  hypothesis is yet agreed upon.

45
Revisiting genetic versus environmental influences

• A classic experiment that opened many eyes.

46
Clausen Keck and Hiesey 1948

• Worked with Achillea lanulosa
• On average plants from the low altitude
  Populations produce slightly more stems than
  those native to higher elevations. (30.20 to
  28.32)

Figure 9.31 p. 354
47
When grown together at low elevation, low
elevation plants produced more stems
This is consistent with the idea that
high-altitude plants are genetically programmed
to produce fewer stems
48
When the two source plants were grown together at
high altitude .

• High altitude plants had more stems! (19.89 vs
  28.32)
• Each population was superior in its own
  environment
• Apparently there are genetic differences that
  control how each responds to the environment
• This is a demonstration ofphenotypic plasticity

49
The upper part of the illustration shows the
different appearances of Achillea
lanulosa-populations like the variance in the
plants' heights cultivated under identical
standard conditions in climatic chambers. The
lower part of the illustration gives the natural
geographic origin of the single populations by
way of a profile of a west-to-east cross-section
through California to the left Sierra Nevada,
to the right Great Basin (J. CLAUSEN, D. D.
KECK, W. M. HIESEY, 1948)
50
Limits of heritability studies

• Must always remember that variation has both a
  genetic and an environmental component.
• Any estimate of heritability is specific to a
  particular population living in a particular
  environment.
• High heritability within groups tell us nothing
  about the origin of the differences between
  groups
• Cannot be used to determine the differences
  between populations of the same species that live
  in different environments.

51

• All that we can really gain by measuring
  heritability is the ability to predict whether
  selection on the trait will cause a population to
  evolve

52
The End