Exercise 15.2

Breaking Genetic Correlations

(This exercise is based on Delph, L.F., J.C. Steven, I. A. Anderson, C. R. Herlihy, and E. D. Brodie III. 2011. Elimination of a genetic correlation between the sexes via artificial correlational selection. Evolution 65-10: 2872–2880)

(Note: The reference above links directly to the article on the journal’s website. In order to access the full text of the article, you may need to be on your institution’s network [or logged in remotely], so that you can use your institution’s access privileges.)

INTRODUCTION

A key question in evolutionary developmental biology is the extent to which genetic correlations constrain the evolution of traits. A related question is how rapidly genetic correlations can be broken. These questions are particularly important when dealing with traits that differ between males and females.

The optimal trait values differ markedly between males and females (e.g., large flowers may be favored in females but disfavored in males). Yet, genes that influence male function often substantially overlap with those that influence female function. Thus, strong genetic correlations often exist between female and male traits, and selection on female traits can produce correlated responses in male traits, and vice versa.

A team of Indiana University researchers, led by Lynda Delph, examined whether artificial selection could reduce genetic correlations between male and female flowers. Their study organism was Silene latifolia, or white campion, a dioecious plant (individuals have either male organs or female organs, but not both).

In white campion, females typically have flowers with much wider calyxes (sepals). There is a strong positive genetic correlation between male and female calyx width: The sisters of males that have flowers with a wider-than-average calyx typically have flowers with a wider-than-average calyx.

Delph and colleagues used artificial selection to manipulate the correlation between male collected data on male and female flower sizes. They had two sets of three lines each: control and experimental. In the experimental lines, they used family selection, wherein they took the mean of the calyx width for male and female flowers. For these lines, they then proceeded to select based on the following diagram. In the control lines, individuals were chosen at random to start the next generation.

QUESTIONS

Use the information in Figure 1 to answer questions 1 through 3.

Figure 1 For each family, trait values are plotted in a two-dimensional grid. The x-axis shows the mean calyx widths of female flowers, and the y-axis shows the mean calyx widths of male flowers. The diagonal line without arrows shows the regression line. This is also the major axis line. Perpendicular to the regression line (with arrows) is the minor axis line.

Question 1. Why must family selection, not individual selection, be used when trying to change the genetic correlation between male and female traits?

Question 2. Is the correlation between male and female flower calyx width in the figure above positive or negative, strong or weak?

Question 3. What do you notice about the selected families in relation to the regression/major axis line? Relative to the non-selected families, what can be said about the mean male calyx widths of male flowers in the selected families?

Use the information in Figure 2 to answer questions 4 through 8.

Figure 2 The figure above shows the response to selection on the correlation between the female and male traits. In the line graph, the correlation is on the y-axis and the time (in generations) is on the x-axis. Points marked “∗”are significantly less than 1 but not significantly greater than 0. Red points and lines designate the control lines, while blue points and lines designate the experimental lines. The point marked “†” is significantly less than 1 and significantly greater than 0. All other points are not significantly less than 1. The inserts show the raw data for specific lines, with female-trait family means on the x-axis and male-trait family means on the y-axis.

Question 4. What happens to the correlation of male and female traits in selection line 1 over the course of artificial selection? Does the correlation respond to selection?

Question 5. What happens to the correlation of male and female traits in selection lines 2 and 3 over the course of artificial selection? Does the correlation respond to selection?

Question 6. Describe the pattern of female- and male-trait values in generation 5 in selection line 3.

Question 7. After running the selection to alter the correlation between male and female traits, the researchers wanted to observe how the lines would respond to ordinary directional selection on female traits. Suppose there is a strong positive selection to reduce the size of female flowers, and the genetic correlation between male and female flower size is positive and close to 1. Based on this, what do you expect would happen to the male flower size?

Question 8. Suppose there is a strong positive selection to reduce the size of female flowers, and the genetic correlation between male and female flower size is close to 0. Based on this, what do you expect would happen to the male flower size?

Use the information in Figure 3 to answer question 9.

Figure 3 Percentage of female-trait change (F) and male-trait change (M) in response to directional selection to reduce female flower size. Delph and colleagues selected for reduced female flower size (calyx width) in both C1 and S3 lines. In C1, the genetic correlation between female and male traits is positive and close to 1. In S3, the genetic correlation between female and male traits is close to 0.

Question 9. Describe the observations. Do these fit with your expectations from Questions 7 and 8?