Phenotypic Correlation

1
Phenotypic Correlation

• degree to which two traits co-vary among
  individuals in a population

lifespan wing width
wing length thorax width
2
Genetic Correlation

• phenotypic correlation genetic environmental
  correlation
• rP rA rD rE
• rP hxhyrA
  exeyrE

3
To measure genetic correlations

• Use same breeding designs as for additive
  variance, but measure two traits instead of just
  one
• Correlations between traits in parents and
  offspring
• Correlation between traits in half-sibs and
  full-sibs
• OR
• Perform artificial selection

4
Why are traits genetically correlated?

• Pleiotropy single gene affects multiple traits
• in morning glories, floral pigment also affects
  plant toxicity to herbivores
• in fruit flies, a mutation in insulin receptor
  doubles lifespan but causes sterility

5

• Pin and Thrum morphs in Primrose
• Homostylus flowers are rare but do occur.
• Linkage disequilibrium nonrandom association
  between alleles at two loci
• Selection for particular combination of alleles
  at two loci generate a strong genetic
  correlation.
• Close linkage necessary

6
Genetic correlations in natural populations
7
Natural selection and evolution in correlated
traits

• Selection on one trait will lead to correlated
  evolution in the other trait
• Traits that are negatively correlated are said to
  exhibit tradeoffs.
• Offspring size and number
• Fecundity and lifespan
• Development rate and adult size

8
Selection experiment
9
Measuring selection

• On phenotypes (mostly morphology)

10
Visualizing selection on a morphological trait
11
Linear regression y a bxw a bz
Relative fitness
Phenotype (wing length)
12
Simple fitness functions
13
Linear and nonlinear terms y a bx cx2 w
a bz gz2
Relative fitness
Phenotype (wing length)
14
a bz gz2If g is negative,slope is
decreasing if g is positive, slope is
increasing.b measures linear selection. g
measures nonlinear (or quadratic) selection.
15
Real fitness functions
Wild radish
16
Selection with correlated traits
Which trait is the target of selection?
17
Multiple regression y a b1x1 b2x2 w
a b1trait1 b2trait2 b3trait3
partial regression coefficients
18
Total selection on a trait is sum of direct
selection indirect selection on correlated
traits
S1 b1 b2r12 b3r13
19
Experimental manipulation to measure selection

• Statistical measures cannot prove the existence
  of direct selection on a trait.
• Why?
• But statistical measures can indicate which
  traits are probably not targets of direct
  selection and which traits warrant further study.
• To prove selection on a trait, experimental
  manipulation is necessary. This is the only way
  to eliminate possibility of correlation with
  unmeasured traits.

20
Experimental approaches

• Directly manipulate the trait you think is under
  direct selection
• artificially lengthen and shorten the horns
  of fungus beetles
• Artificially manipulate color, pattern, display
  sites (e.g., bowers),

21
Manipulation of morphology
(Andersson 1982)
22
Things to consider before doing such an experiment

• How much variation should you create (beyond the
  range of variation in the current population)?
• Does the manipulation itself affect performance
  (can the birds fly with manipulated tails)?
• What if you cant do a direct manipulation (finch
  beak size)?

23
Plasticity in plant height as an adaptation to
density

• Impatiens capensis grows taller when grown at
  high density. Is this an adaptation, or a
  deleterious consequence of suboptimal conditions?
• Grow plants to different heights in the
  greenhouse by using different ratios of redfar
  red light.
• Transplant tall and short plants to the field in
  both low and high density.

(Dudley and Schmidt 1982)
24
Plasticity in plant height as an adaptation to
density

• Selection for height was positive at high
  density. bgt0 indicating selection on increased
  height at high density
• But b lt0 at low density, indicating selection to
  decrease height at low density.
• Interpretation tall plants better able to
  compete for light under crowded conditions, but
  there is a fitness cost to being tall when light
  is not limiting

(Dudley and Schmidt 1982)