Evolution of Polygenic Traits


NOTE: These are lecture notes for Biology 391, Organic Evolution, at The University of Tennesee at Martin.  Anyone outside of UT Martin wishing to use these notes or to contact me for additional information should first read the information obtained by clicking here.

Goals: the goals of this lecture are to evaluate how polygenic traits evolve and compare patterns of evolution of polygenic traits to patterns of evolution of single gene traits.

Related Textbook Material: Freeman and Herron (2001) Chapter 7

Lab Manual Questions over this material are in Lab Manual Chapter IX

The Lecture:

You have learned, or are learning, from your computer modeling assignment, the patterns of evolutionary change in traits coded by a single gene. In this lecture, you will learn about the patterns of evolutionary change in the traits coded by many genes, quantitative or polygenic traits, which were introduced in the last lecture.

Remember that the distribution of individuals with different trait values for polygenic (quantitative) traits in a population is typically a bell-shaped curve, as shown here:

There are three main ways selection could act on a population, given a distribution of traits such as this. These are:

  1. Directional selection: the situation in which one extreme form of the trait has highest fitness.
  2. Stabilizing selection: the situation in which the average form of the trait has higher fitness than does either extreme.
  3. Disruptive selection: the situation in which both extreme forms of the trait have higher fitness than does the average.
These forms of selection, and the expected results of these forms of selection over time, are illustrated in the following figure:

The results of selection on quantitative traits generally makes sense -- the forms that have highest fitness become most common. As shown, directional selection results in a change in the mean value of the trait toward the form that has highest fitness. Stabilizing selection results in the loss of the extreme forms of the trait; this means there is a decrease in genetic variation -- eventually, genetic variation may be lost, as all individuals will have the alleles for the highest fitness, average, trait value. At this point, any phenotypic variation would depend on direct environmental effects rather than on genetic differences among individuals, and the heritability of the trait would be zero, or at least very low. Disruptive selection results in an increase in both extremes and a loss of intermediate forms.

Over a long period of time directional selection will result in a shift in the frequency of individuals with different traits until the average form has highest fitness. At this point, the situation becomes one of stabilizing selection, and the extreme forms of the trait will be lost. So directional selection eventually will lead to a situation where genetic variation will be lost (heritability will become zero) and all individuals will have the alleles for the highest fitness form of the trait.

Disruptive selection is somewhat more complicated. If it is strong enough, it could lead to speciation -- the groups of individuals with different extremes of the trait could evolve to be so different that they can no longer reproduce. We will consider under what circumstances this might happen when we study speciation later in the semester. If speciation does occur, we'd expect each group to evolve until all individuals have the alleles for the highest form of the trait, so that within each group genetic variation will be lost (heritability will become zero.)

Look over these forms of selection again and note that the typical ultimate response is a loss of genetic variation as all individuals end up with the alleles for the highest fitness form of the trait. This is a typical result of natural selection. Compare this to what you have discovered about single gene traits from your computer modeling assignment and you will see that in most single gene situations natural selection also results in loss, or near loss, of genetic variation. You should discover just one exception to this in your computer modeling exercise: the exception, for single gene traits, is heterosis. Heterosis maintains genetic variation because the heterozygote has highest fitness and has both alleles, so reproduction of heterozygotes keeps both alleles (genetic variation) present in the population.

Superficially, heterosis seems similar to stabilizing selection in that in both heterosis and stabilizing selection the intermediate form has higher fitness than does either extreme. The difference is that heterosis applies to single gene traits, and the intermediate form is the heterozygote, which is produced when an individual has two different alleles for the trait. Stabilizing selection applies to quantitative traits, and the intermediate form is not necessarily a heterozygote for any of the genes that make up the traits. Consider different ways in which an individual could have the intermediate phenotype (trait value) for a quantitative trait:

To see the different ways in which an individual could have an intermediate phenotype for a quantitative trait, consider a quantitative trait that is the size of some part of the body of an individual. Suppose some alleles tend to make the individual a little larger; we'll call these "+" alleles. Suppose some other alleles tend to make the individual a little smaller; we'll call these "-" alleles. To have an intermediate phenotype, an individual should have half "+" and half "-" alleles. Since the trait is polygenic, there are many genes where there could be "+" and/or "-" alleles.

One way in which an individual could have the intermediate phenotype is to be heterozygous (+/-) at each gene that affects this trait. The other way in which an individual could have the intermediate phenotype is to be homozygous "+,+" at half of the genes that affect the trait, and homozygous "-,-" at the other half of the genes that affect the trait. Both of these ways result in the same phenotype of an individual because in both cases half the alleles that affect the trait are "+" and half are "-"

When stabilizing selection is occurring, the intermediate phenotype has highest fitness, so individuals that have the intermediate phenotype in either of the genetic ways described in the previous paragraph will have high survival. The interesting thing, though, is that they will differ in the survival of their offspring. Individuals who are heterozygous "+,-" for all genes can pass on a wide variety of different gametes -- each gene can contribute either a "+" or a "-" allele to the gamete. As a result, the gamete could turn out to have all "-" alleles, all "+" alleles, or some combination. Offspring produced by gametes that are all "-" will tend to be small; offspring produced by gametes that are all "+" will tend to be large. Since selection is stabilizing, the extremes (small and large) have low fitness, so these offspring will have lower survival. Some offspring will get mixtures of "+" and "-" alleles and will have higher survival. The point is that when individuals who are intermediate because they are heterozygous reproduce they produce offspring with a lot of variation in the phenotype, and the extreme forms will not survival.

Contrast the above situation with the situation when individuals are intermediate because they are homozygous "+,+" at half their genes and homozygous "-,-" at the other half of their genes reproduce. The genes that are "+,+" can contribute only "+" alleles to the gametes; the genes that are "-,-" at half their genes can contribute only "-" alleles. So the gamete will have half "+" and half "-" alleles. Offspring of such individuals are more likely to have the intermediate phenotype, and therefore high fitness. There is much less variation in offspring phenotype, and therefore offspring survival, of such individuals.

The result of this is that individuals who are intermediate because they are homozygous "+,+" at half their genes and homozygous "-,-" at the other half of the genes affecting a trait will have more surviving offspring, and more grandchildren, than do individuals who are intermediate by being heterozygous. Over time, as a result, all individuals will end up homozygous, and at this point the alleles will be fixed for all genes -- there will be no more genetic variation in the population, so the heritability of the trait will be zero.

Now that we've seen this rather long explanation for why stabilizing selection results in loss of genetic variation, remember again that the forms of natural selection we've seen so far on both single gene and polygenic traits almost all (except for heterosis) result in a great decrease or complete loss of genetic variation. This information has been summarized by a statement called Fisher's Fundamental Theorem of Natural Selection, which states that traits subject to strong natural selection rapidly lose genetic variation. Strong natural selection means that there are large differences in fitness between the different forms of the trait.

Think about this point. Natural selection, as you saw in the first lecture, only occurs if genetic variation is present. The effect of natural selection, however, is to decrease genetic variation. You can see, therefore, that the effect of natural selection occurring could be to make further natural selection less likely.

Since natural selection decreases genetic variation, we would expect low heritability (since heritability is a measure of genetic variation in a population) in many quantitative traits. This is observed in some traits in natural populations, but there are also examples, such as the Darwin's Finches discussed in your textbook, in which there is genetic variation in quantitative traits. There are some situations in which genetic variation will be maintained over time. We will look at three of these situations in this lecture, since these apply specifically to quantitative (polygenic) traits. Later, we will consider some other situations in which genetic variation will be maintained over time.

  1. One situation in which genetic variation may be maintained over time in a quantitative trait occurs if stabilizing selection occurs, but is fairly weak. Weak selection means that while individuals with some traits do survive and reproduce better than others, the differences in survival and reproduction of individuals with different traits are small. In this situation, loss of genetic variation will be slow, and the loss of variation through selection may be balanced by the creation of genetic variation that occurs through mutation. Since mutation is such a slow process, it would not maintain significant variation for a single gene trait, but for a polygenic trait for which mutation at any of many different genes can add genetic variation to a trait, it is at least theoretically possible for the increase in genetic variation through mutation to balance the loss of variation through weak stabilizing selection. This situation has been shown to work in theoretical models but is very hard to test, so we do not know how often it occurs.
  2. A second situation in which genetic variation may be maintained over time in a quantitative trait occurs if there is directional selection, but the environment changes fairly rapidly so that in some years one extreme has highest fitness (for example, largest individuals have high fitness) and in other years the other extreme has highest fitness (for example, smallest individuals have high fitness.) In this case, the environment changes too rapidly for there to be time for genetic variation to be lost. An example of this occurs in Darwin's Finches in the Galapagos Islands. During most years, the climate on the islands is extremely dry (they are desert islands.) The plants that reproduce during those years have thick seeds, and finches (seed-eating birds) with large, thick bills are best able to crack and eat these seeds, so having a thick bill has highest fitness during dry years. Every 5 years or so there is what is called an El Nino year, and during these years it rains for months on end. The plants that reproduce during these years produce many small seeds. Finches with small, thin bills can easily eat these seeds and since these finches are also typically smaller in body size and mature earlier so they reproduce more quickly than larger birds. As a result, during these very wet years, finches with small bills have highest fitness. As discussed in your textbook, bill size in Darwin's Finches has high heritability, meaning there is a lot of genetic variation. This variation is maintained because of these changes in the environment that change the extreme that has highest fitness under directional selection.
  3. A third situation in which genetic variation may be maintained in quantitative traits occurs if one extreme of a trait has higher fitness at one point in the life or for one reason, but the opposite extreme of the trait has higher fitness in a different context. The balance of these different effects of selection may result in a variety of intermediate forms that all have about the same fitness, so all are maintained in the population over time. An example of this occurs in another trait of Darwin's Finches. In these finches, older males are black and females are brown. One-year-old males vary in plumage color. Some are as black as older males, some are brown like females, and most are in between -- some with more black feathers and some with more brown feathers. Mid-parent offspring regressions have shown that plumage color variation in one-year old males has high heritability, meaning that the differences in plumage color between these birds are genetic. It turns out that brown males are less likely to attract females, so for that reason we'd expect black males to have higher fitness. Black males, however, attract more aggressive attacks from adult males than do brown males, so brown males survive better and for that reason we'd expect brown males to have higher fitness. The balance between these effects is probably what maintains a mixture of black and brown males in the population.
At this point, you should be able to describe the effects of the main different forms of selection (directional, stabilizing, and disruptive) on polygenic traits on the average trait value AND on the amount of genetic variation (which we would measure as heritability) in the trait. You should compare and contrast these effects with the effects of the different forms of selection you've studied for single gene traits with your computer modeling assignment. You should see that most of these forms of selection will result in a loss of genetic variation (low heritability.) You should be able to explain why stabilizing selection results in loss of genetic variation. Finally, you should learn the exceptions we've discussed to the general rule (Fisher's Fundamental Theorem) that natural selection results in loss of genetic variation. These exceptions are: heterosis (this applies only to single-gene traits), a balance between mutation and weak stabilizing selection, directional selection in an environment that changes so that the extreme trait value that has highest fitness fluctuates back and forth over time, and a balance between different forms of selection affecting the same trait. Based on this lecture and the previous lecture, you should now try to answer all the questions in Lab Manual Chapter X.

Click here to return to the index of lectures