1 of 55

Evolutionary Analysis

Fourth Edition

Chapter 9

Evolution at Multiple Loci:

Quantitative Genetics

Scott Freeman • Jon C. Herron

2 of 55

Japanese flounder do not fall into discrete color categories. Instead, they show continuous variation.

3 of 55

The histogram below shows that color intensity among these fish is normally distributed

4 of 55

Quantitative Genetics

5 of 55

9.1 The Nature of Quantitative Trait

Qualitative Traits vs. Quantitative Traits

Quantitative traits are traits for which the distribution of phenotypes is continuous rather than discrete.

6 of 55

Figure 9.1 Some quantitative traits in humans

7 of 55

8 of 55

9 of 55

10 of 55

Figure 9.2 Mendelian genetics can explain quantitative traits

11 of 55

Quantitative traits

Quantitative traits are consistent with Medelian genetics. They are influenced by the combined effects of the genotype at many loci.

12 of 55

Figure 9.3 Edward East's data confirm the predictions of the Mendelian model in Figure 9.2c

13 of 55

Figure 9.4 Quantitative traits are influenced by the environment as well as genotype.

These three yarrow plants were grown from cuttings of the same individual, and are thus genetically identical. Reared at different altitudes, they show dramatic differences in height.

14 of 55

9.2 Identifying Loci That Contribute to Quantitative Traits

QTLs (Quantitative Trait Loci): The loci that influence quantitative traits.

15 of 55

QTL Mapping

16 of 55

Figure 9.5 A phylogeny of Mimulus cardinalis, Mimulus lewisii, and kin. The common ancestor of these species was pollinated by bees. Pollination by hummingbirds evolved twice: once in the common ancestor of M. eastwoodiae and kin, and once in M. cardinalis.

17 of 55

Figure 9.6 Mimulus cardinalis, Mimulus lewisii, and their F₁and F₂descendents

Photo (a) shows M. lewisii, photo (b) shows an F₁ hybrid, and photo (c) shows M. cardinalis. The remaining photos (d-l) show F₂ hybrids produced by crosses between F₁s.

18 of 55

QTL Mapping

We can detect the presence and location of loci influencing a quantitative trait by crossing parents from populations with fixed difference.
Among the grand-offspring, we look for association between phenotype and genotype at marker loci.

19 of 55

20 of 55

Figure 9.7 The logic of QTL mapping

F2 populations

21 of 55

Figure 9.9 QTLs for floral traits in Mimulus lewisii and Mimulus cardinalis, sorted by the strength of their effects on the phenotype

Bradshaw and colleagues (1998) found between one and six QTLs for each of the dozen floral traits they mapped. Differences in genotype for the majority of these QTLs explained only a modest amount of the variation in flower phenotype, as indicated by the higher bars on the left side of these graphs. However, for 9 of the 12 floral traits, there was at least one QTL at which differences in genotype explained more than 25% of the variation in phenotype. These QTLs with large phenotypic effects are indicated by the bars in the center and right of these graphs. There are separate graphs for M. lewisii and M. cardinalis because most of the marker loci used by Bradshaw and colleagues can only be detected in one species or the other. However, the QTLs themselves almost certainly occur in both genomes. Rerendered from Bradshaw and colleagues (1998).

22 of 55

Candidate Loci

We can confirm that a particular locus influences a quantitative trait by looking for associations between genotype and phenotype.

23 of 55

Figure 9.10 A novel allele at a single locus can dramatically alter a flower's attractiveness to different pollinators

24 of 55

Figure 9.11 Identification of a quantitative trait locus influencing a personality trait in human.

Sequence variation at the D4 dopamine receptor locus can be reduced to two categories of alleles: short (S) and long (L).

25 of 55

9.3 Measuring Heritable Variation

Quantitative genetics allows us to analyze evolution by natural selection in traits controlled by many loci.

26 of 55

Heritable Variation

The first step in a quantitative genetic analysis is to determine the extent to which the trait in question is heritable.
That is, we must partition the total phenotypic variation (V_P) into a component due to genetic variation (V_G) and a component due to environmental variation (V_E).

27 of 55

Heritability

Heritability = V_G/V_P = V_G/(V_G+V_E)

(broad-sense heritability)

28 of 55

Figure 9.13 Scatterplots showing offspring height as a function of parent height

29 of 55

Narrow-sense Heritability

The slope represents a version of the heritability h²

The heritability h² , is a measure of the , V_A in a trait.

h²V_A/V_P = V_A/(V_A+V_D+V_E) (V_D, dominance genetic variation)

(additive) genetic variation

30 of 55

Figure 9.14 A field experiment on the heritability of beak size in song sparrows

31 of 55

Figure 9.16 Estimating heritability from twin studies

Monozygotic twins develop from a single zygote, and thus share all their genes. Dizygotic twins develop from separate zygotes, and share half their genes. If the heritability of a trait is high, monozygotic twins will resemble each other more strongly than dizygotic twins.

32 of 55

9.4 Measuring Differences in Survival and Reproductive Success (Selection Strength)

33 of 55

The Strength of Selection

The 2^nd step in a quantitative genetic analysis is to measure the strength of selection on the trait in question.

34 of 55

The Strength of Selection

One measure is the selection differential, S = the difference between the mean of the selected individuals and the mean of the entire populations.
A second (and related) measure of the strength of selection is the selection gradient.

35 of 55

Figure 9.17 Measuring the strength of selection

36 of 55

Selection Gradient 𝞫

The selection gradient, 𝞫, for trait t is equal to the selection differential, S, divided by the variance:
𝞫 = S/var(t)

37 of 55

Figure 9.18 The response to selection R, is equal to the heritability multiplied by the selection differential

38 of 55

Alpine Skypilots and Bumblebees

39 of 55

40 of 55

41 of 55

Figure 9.20 Estimating the heritability of flower size (corolla flare) in alpine skypilots

This scatterplot shows offspring corolla flare as a function of maternal plant corolla flare for 58 skypilots. The slope of the best-fit line is 0.5

42 of 55

Figure 9.21 Estimating the selection gradient in alpine skypilots pollinated by bumblebees

This scatterplot shows relative fitness (number of surviving 6-year-old offspring divided by average number of surviving 6-year-old offspring) as a function of maternal flower size (corolla flare). The slope of the best-fit line is 0.13.

43 of 55

Figure 9.22 Measuring the evolutionary response to selection in alpine skypilots

These histograms show the distribution of flower size (corolla flare) in the offspring of hand-pollinated skypilots (a; average = 13.1 mm) and bumblebee-pollinated skypilots (b; average = 14.4 mm).

44 of 55

9.6 Modes of Selection and the Maintenance of Genetic Variation

45 of 55

46 of 55

Modes of Selection

Directional selection and stabilizing selection tend to reduce the amount of variation in a population;
Disruptive selection tends to increase the amount of variation.

47 of 55

Figure 9.26 Stabilizing selection on a gall-making fly

(a) Parasitoid wasps kill fly larvae inside small galls at higher rates than they kill larvae inside large galls. (b) Birds kill fly larvae inside large galls at higher rates than they kill larvae inside small galls. (c) The distribution of gall sizes before (tan + red portion of bars) and after (red portion of bars) selection by parasitoids and birds. Overall, fly larvae inside medium-sized galls survived at the highest rates.

48 of 55

Figure 9.27 Disruptive selection on bill size in the black-bellied seedcracker (Pyrenestes o. ostrinus)

49 of 55

9.7 The Bell-Curve Fallacy and Other Misinterpretations of Heritability

50 of 55

Figure 9.28 High heritability within populations tells us nothing about the cause of differences between populations

The plants in the Stanford population are taller, on average, than the plants in the Mather population. We know these two populations are genetically identical because they were grown from cuttings of the same seven plants.

51 of 55

Common Garden Experiments�

The only way to determine the cause of differences between populations is to rear individuals from each of the populations in identical environments.

52 of 55

Common Garden Experiments

Figure 9.31a Data from experiments by Clausen, Keck, and Hiesey (1948)

53 of 55

Garden at low altitude

54 of 55

Garden at high altitude

55 of 55

End