Artificial Selection

1
Artificial Selection

• Introduction
• Breeders equation
• Short-term response to selection
• Long-term response to selection

2
Artificial selection introduction

• Artificial selection, in the form of fancy
  pigeons, was Darwins inspiration for the theory
  of natural selection
• Directly analogous to natural selection
• Fitness largely defined by experimenter
• Darwinian fitness still involved

3
Artificial selection deterministic change in
allele frequencies

• Effect of selection depends on s and the initial
  frequency of q.

4
Dominance and selection

• Dominance is not heritable (individual alleles
  inherited from each parent, not pairwise
  combinations)
• However, dominance clearly plays a role in
  selection

5
Dominance and Selection I
Dq

• Change of gene frequency with s 0.2, for
  different values of initial gene frequency q.
• Upper panel no dominance lower panel, complete
  dominance
• Selection against q
• Selection for q

Dq
q
6
Dominance and Selection II
q

• Change of gene frequency during the course of
  selection from one extreme the other, with s
  0.2.
• Upper panel no dominance lower panel, complete
  dominance
• Selection against q
• Selection for q

q2
Generations
7
Additive variation

• Because alleles are inherited singly, artificial
  selection concerns itself with additive
  variation, thus h2.

8
Artificial selection is

• the deliberate choice of a select group of
  animals or plants, usually superior for a trait
  or traits, for breeding
• Three major outcomes of artificial selection
• Change in mean of selected trait
• Change in variance of selected trait
• Change in traits covarying with selected trait

9
Artificial Selection

• Introduction
• Breeders equation
• Short-term response to selection
• Long-term response to selection

10
The breeders equation

• Consider a random mating population. Plot
  offspring y vs. midparent x in this population.

11
The breeders equation

• R mean difference between offspring and
  population before selection
• S mean difference between selected parents and
  population before selection
• R ,S lies on the parent-offspring regression

R
S
12
The breeders equation I

• R ,S is the mean of the selected parents and
  their offspring
• R/S is the slope of the line, the
  parent-offspring regression
• Thus
• Or, R h2S

13
The breeders equation II
14
The breeders equation II
15
The breeders equation II
16
The breeders equation II
17
The breeders equation II
18
Why is offspring mean always less than parent
mean?

• Because (a trivial explanation)
• So why not
• Alleles, not genotypes, are inherited
• Dominance and epistasis
• Good phenotype can be result of environment

19
What is required for the breeders equation to
work?

• Good estimate of h2
• Standardize effects of environment between
  generations
• No change in allele frequencies between
  generations (natural selection, mutation, drift)

20
Factors that influence S

• Proportion selected
• High proportion, S small
• Low proportion, S large

21
Factors that influence S

• Phenotypic variance
• VP large, S small
• VP small, S large

22
Dimensionless variables

• We can compare responses of different traits,
  organisms if we avoid using phenotypic units (mm,
  kg, days, etc.)
• Standardize to R and S to the phenotypic standard
  deviation

23
Dimensionless variables II

• Divide through by the phenotypic standard
  deviation
• the intensity of selection

24
Improving the response to selection

• Increase h2
• Decrease environmental variance
• Decrease measurement error
• Increase i
• However, risk increasing inbreeding

25
Measurement of response to selection

• Response seldom linear
• Stagger due to drift, environmental fluctuation,
  sampling error, differences in i
• Thus need several generations to measure R well

24 22 20 18 16 14
Generation mean, in grams
2 4 6 8 10
Generations
26
Environmental fluctuation

• Control statistically by maintaining contemporary
  unselected (control) or divergently selected
  population
• Environmental fluctuation should affect selected
  and control (or divergently selected) lines
  equally

27
Artificial Selection

• Introduction
• Breeders equation
• Short-term response to selection
• Long-term response to selection

28
Replication of selected lines

• Multiple reps. of control, selected lines are
  kept to control for genetic drift
• Multiple reps. essential for study of correlated
  response to selection
• Repeatability of reps. high Vrep many low
  frequency alleles in base pop.

29
Realized heritability

• Realized h2 cumulative selection
  response/cumulative selection differential
• Best measurement of effectiveness of selection
• Comparable between experiments/traits
• Best predictive nature

30
Symmetry of selection response

• Requires up and down selected lines
• Asymmetry can give information about constraints
  on the trait

31
Causes of asymmetry in selection response

• Selection constraint natural selection could
  enhance selection in one direction, inhibit it in
  the other
• Chance control by using multiple reps.
• Inbreeding depression mean decreases with
  inbreeding depression, therefore down selection
  appears more effective
• Extreme initial allele frequencies in base
  population

32
Artificial Selection

• Introduction
• Breeders equation
• Short-term response to selection
• Long-term response to selection

33
Long-term response is hard to predict

• Theoretically, in a finite population with no
  mutation, response to selection should be finite
  eventually all up alleles will be fixed in the
  up line, etc.
• However, new mutations occur continually,
  therefore response should occur indefinitelybut
  slowly

34
Generalities about long-term response

• Response does continue for a long time 20-30
  generations
• Response to selection in the range of 10-20 i.e.
  generating individuals far more extreme than ever
  seen in base population

35
Response to selection does not proceed
indefinitely I

• Mutations occur too rarely
• Mutations pushing trait beyond extreme are too
  deleterious (constraint)

36
Response to selection does not proceed
indefinitely II

• Dominance variation
• Overdominance (heterozygosity required for
  phenotype, equilibrium allele frequency reached)
• Negative pleiotropy