Tải bản đầy đủ
3: Heritability Is Used to Estimate the Proportion of Variation in a Trait That Is Genetic
Question: How is flower length in tobacco plants inherited?
31 34 37 40 43 46
x = 40.5 mm
84 87 90 93 96 99 102
x = 93.3 mm
1 Flower length in the F1 was
about halfway between that
in the two parents,…
55 58 61 64 67 70 73
2 …and the variance in the F1
was similar to that seen in the
3 The mean of the F2 was similar
to that observed for the F1,…
52 55 58 61 64 67 70 73 76 79 82 85 88
Flower length (mm)
4 …but the variance in the F2
was greater, indicating the
presence of different genotypes among the F2 progeny.
Conclusion: Flower length of the F1 and F2 is consistent
with the hypothesis that the characteristic is determined by
several genes that are additive in their effects.
16.10 Edward East conducted an early statistical study of
the inheritance of flower length in tobacco.
Consider a dairy farmer who owns several hundred milk
cows. The farmer notices that some cows consistently produce more milk than others. The nature of these differences
is important to the profitability of his dairy operation. If the
differences in milk production are largely genetic in origin,
then the farmer may be able to boost milk production by
selectively breeding the cows that produce the most milk. On
the other hand, if the differences are largely environmental
in origin, selective breeding will have little effect on milk
production, and the farmer might better boost milk production by adjusting the environmental factors associated with
higher milk production. To determine the extent of genetic
and environmental influences on variation in a characteristic, phenotypic variation in the characteristic must be partitioned into components attributable to different factors.
To determine how much of phenotypic differences in a population is due to genetic and environmental factors, we must
first have some quantitative measure of the phenotype under
consideration. Consider a population of wild plants that differ in size. We could collect a representative sample of plants
from the population, weigh each plant in the sample, and
calculate the mean and variance of plant weight. This phenotypic variance is represented by VP.
Components of phenotypic variance First, some of the
phenotypic variance may be due to differences in genotypes
among individual members of the population. These differences are termed the genetic variance and are represented
Second, some of the differences in phenotype may be
due to environmental differences among the plants; these
differences are termed the environmental variance, VE.
Environmental variance includes differences in environmental factors such as the amount of light or water that the plant
receives; it also includes random differences in development
that cannot be attributed to any specific factor. Any variation
in phenotype that is not inherited is, by definition, a part of
the environmental variance.
Third, genetic–environmental interaction variance
(VGE) arises when the effect of a gene depends on the specific environment in which it is found. An example is shown
in Figure 16.11. In a dry environment, genotype AA produces a plant that averages 12 g in weight, and genotype aa
produces a smaller plant that averages 10 g. In a wet environment, genotype aa produces the larger plant, averaging 24 g
in weight, whereas genotype AA produces a plant that averages 20 g. In this example, there are clearly differences in the
two environments: both genotypes produce heavier plants in
the wet environment. There are also differences in the
weights of the two genotypes, but the relative performances
of the genotypes depend on whether the plants are grown in
a wet or a dry environment. In this case, the influences on
16.11 Genetic–environmental interaction variance is
obtained when the effect of a gene depends on the specific
environment in which it is found. In this example, the genotype
affects plant weight, but the environmental conditions determine
which genotype produces the heavier plant.
variance primarily determines the resemblance between parents and offspring. For example, if all of the phenotypic variance is due to additive genetic variance, then the phenotypes
of the offspring will be exactly intermediate between those of
the parents, but, if some genes have dominance, then offspring may be phenotypically different from both parents
(i.e., Aa ϫ Aa S aa offspring).
Second, there is dominance genetic variance (VD)
when some genes have a dominance component. In this case,
the alleles at a locus are not additive; rather, the effect of an
allele depends on the identity of the other allele at that locus.
For example, with a dominant allele (T), genotypes TT and
Tt have the same phenotype. Here, we cannot simply add the
effects of the alleles together, because the effect of the small
t allele is masked by the presence of the large T allele. Instead,
we must add a component (VD) to the genetic variance to
account for the way in which alleles interact.
Third, genes at different loci may interact in the same
way that alleles at the same locus interact. When this genic
interaction takes place, the effects of the genes are not additive. For example, Chapter 4 described how coat color in
Labrador retrievers exhibits genic interaction; genotypes BB
ee and bb ee both produce yellow dogs because the effect of
alleles at the B locus are masked when ee alleles are present
at the E locus. With genic interaction, we must include a
third component, called genic interaction variance (VI), to
the genetic variance:
VG ϭ VA ϩ VD ϩ VI
phenotype cannot be neatly allocated into genetic and environmental components, because the expression of the genotype depends on the environment in which the plant grows.
The phenotypic variance must therefore include a component that accounts for the way in which genetic and environmental factors interact.
In summary, the total phenotypic variance can be
apportioned into three components:
VP ϭ VG ϩ VE ϩ VGE
Components of genetic variance Genetic variance can
be further subdivided into components consisting of different types of genetic effects. First, additive genetic variance
(VA) comprises the additive effects of genes on the phenotype, which can be summed to determine the overall effect
on the phenotype. For example, suppose that, in a plant,
allele A1 contributes 2 g in weight and allele A2 contributes
4 g. If the alleles are strictly additive, then the genotypes
would have the following weights:
A1A1 ϭ 2 ϩ 2 ϭ 4 g
A1A2 ϭ 2 ϩ 4 ϭ 6 g
A2A2 ϭ 4 ϩ 4 ϭ 8 g
The genes that Nilsson-Ehle studied, which affect kernel
color in wheat, are additive in this way. The additive genetic
Summary equation We can now integrate these components into one equation to represent all the potential contributions to the phenotypic variance:
VP ϭ VA ϩ VD ϩ VI ϩ VE ϩ VGE
Equation 16.7 provides us with a model that describes the
potential causes of differences that we observe among
individual phenotypes. Importantly, note that this model
deals strictly with the observable differences (variance) in
phenotypes among individual members of a population; it
says nothing about the absolute value of the characteristic
or about the underlying genotypes that produce these
Types of Heritability
The model of phenotypic variance that we’ve just developed
can be used to address the question of how much of the phenotypic variance in a characteristic is due to genetic differences. Broad-sense heritability (H2) represents the
proportion of phenotypic variance that is due to genetic
variance and is calculated by dividing the genetic variance by
the phenotypic variance:
broad-sense heritability = H2 =
The symbol H2 represents broad-sense heritability because it
is a measure of variance, which is in units squared.
Broad-sense heritability can potentially range from 0
to 1. A value of 0 indicates that none of the phenotypic variance results from differences in genotype and all of the differences in phenotype result from environmental variation.
A value of 1 indicates that all of the phenotypic variance
results from differences in genotypes. A heritability value
between 0 and 1 indicates that both genetic and environmental factors influence the phenotypic variance.
Often, we are more interested in the proportion of the
phenotypic variance that results from the additive genetic
variance because, as mentioned earlier, the additive genetic
variance primarily determines the resemblance between parents and offspring. Narrow-sense heritability (h2) is equal
to the additive genetic variance divided by the phenotypic
narrow-sense heritability = h2 =
Having considered the components that contribute to phenotypic variance and having developed a general concept of
heritability, we can ask, How do we go about estimating these
different components and calculating heritability? There are
several ways to measure the heritability of a characteristic.
The mathematical theory that underlies these calculations of
heritability is complex and beyond the scope of this book.
Nevertheless, we can develop a general understanding of
how heritability is measured.
Most methods for calculating heritability compare the
degree of resemblance between related and unrelated individuals or between individuals with different degrees of
relatedness. For example, one method compares the phenotypes of parents and offspring with that of unrelated
individuals. Another method compares the similarity of
identical twins (which have 100% of their genes in common) with the similarity of nonidentical twins (which
have 50% of their genes in common). The general idea is
that, if genes influence genetic differences among individuals, individuals that are more closely related will be more
similar in phenotype. Statistical methods called regression
and correlation are used to determine the degree to which
more closely related individuals are more similar in
An example of calculating heritability by comparing
the phenotypes of parents and offspring is illustrated in
Figure 16.12. Here, the mean phenotype of the parents is
plotted against the mean phenotype of the offspring. Each
data point on the graph represents one family. The line represents the best fit of the all of the points of the graph (deviations of the points from the line are minimized). A
Mean shell breadth of offspring (mm)
Mean shell breadth of parents (mm)
16.12 The heritability of shell breadth in snails can be
determined by plotting the mean phenotype of offspring
against the mean phenotype of the parents. [From L. M. Cook,
Evolution 19:86–94, 1965.]
complex mathematical proof (which we will not go into
here) demonstrates that the slope of the line equals the narrow-sense heritability.
All estimates of heritability depend on the assumption
that the environments of related individuals are not more
similar than those of unrelated individuals. This assumption
is difficult to meet in human studies because related people
are usually reared together. Heritability estimates for
humans should therefore always be viewed with caution.
Broad-sense heritability is the proportion of phenotypic variance
that is due to genetic variance. Narrow-sense heritability is the
proportion of phenotypic variance that is due to additive genetic
variance. Heritability can be measured by comparing the degree of
resemblance between related and unrelated individuals or
between individuals having different degrees of relatedness.
✔ Concept Check 1
If the environmental variance (VE) increases and all other variance
components remain the same, what will the effect be?
a. Broad-sense heritability will decrease.
b. Broad-sense heritability will increase.
c. Narrow-sense heritability will increase.
d. Broad-sense heritability will increase, but narrow-sense
heritability will decrease.
The Limitations of Heritability
Knowledge of heritability has great practical value because it
allows us to statistically predict the phenotypes of offspring on
the basis of their parent’s phenotype. It also provides useful
information about how characteristics will respond to selection (see Section 16.4). In spite of its importance, heritability
is frequently misunderstood. Heritability does not provide
information about an individual’s genes or the environmental
factors that control the development of a characteristic, and it
says nothing about the nature of differences between groups.
This section outlines some limitations and common misconceptions concerning broad- and narrow-sense heritability.
Heritability does not indicate the degree to which a
characteristic is genetically determined Heritability is
the proportion of the phenotypic variance that is due to
genetic variance; it says nothing about the degree to which
genes determine a characteristic. Heritability indicates only the
degree to which genes determine variation in a characteristic.
The determination of a characteristic and the determination of
variation in a characteristic are two very different things.
Consider polydactyly (the presence of extra digits) in
rabbits, which can be caused either by environmental factors
or by a dominant gene. Suppose we have a group of rabbits
all homozygous for a gene that produces the usual numbers
of digits. None of the rabbits in this group carries a gene for
polydactyly, but a few of the rabbits are polydactylous
because of environmental factors. Broad-sense heritability
for polydactyly in this group is zero because there is no
genetic variation for polydactyly; all of the variation is due
to environmental factors. However, it would be incorrect for
us to conclude that genes play no role in determining the
number of digits in rabbits. Indeed, we know that there are
specific genes that can produce extra digits. Heritability indicates nothing about whether genes control the development
of a characteristic; it provides information only about causes
of the variation in a characteristic within a defined group.
An individual does not have heritability Broad- and
narrow-sense heritabilities are statistical values based on the
genetic and phenotypic variances found in a group of individuals. Heritability cannot be calculated for an individual, and
heritability has no meaning for a specific individual. Suppose
we calculate the narrow-sense heritability of adult body weight
for the students in a biology class and obtain a value of 0.6. We
could conclude that 60% of the variation in adult body weight
among the students in this class is determined by additive
genetic variation. We should not, however, conclude that 60%
of any particular student’s body weight is due to additive genes.
There is no universal heritability for a characteristic
The value of heritability for a characteristic is specific for a
given population in a given environment. Recall that broadsense heritability is genetic variance divided by phenotypic
variance. Genetic variance depends on which genes are
present, which often differs between populations. In the
example of polydactyly in rabbits, there were no genes for
polydactyly in the group; so the heritability of the characteristic was zero. A different group of rabbits might contain
many genes for polydactyly, and the heritability of the characteristic might then be high.
Environmental differences may affect heritability because
VP is composed of both genetic and environmental variance.
When the environmental differences that affect a characteristic
differ between two groups, the heritabilities for the two groups
also will often differ. Because heritability is specific to a defined
population in a given environment, it is important not to
extrapolate heritabilities from one population to another.
Even when heritability is high, environmental factors
may influence a characteristic High heritability does
not mean that environmental factors cannot influence the
expression of a characteristic. High heritability indicates
only that the environmental variation to which the population is currently exposed is not responsible for variation in
the characteristic. Let’s look at human height. In most developed countries, heritability of human height is high, indicating that genetic differences are responsible for most of the
variation in height. It would be wrong for us to conclude that
human height cannot be changed by alteration of the environment. Indeed, in several European cities during World
War II, height decreased owing to hunger and disease, and
height can be increased dramatically by the administration
of growth hormone to children. The absence of environmental variation in a characteristic does not mean that the characteristic will not respond to environmental change.
Heritabilities indicate nothing about the nature of
population differences in a characteristic A common
misconception about heritability is that it provides information about population differences in a characteristic. Heritability is specific for a given population in a given
environment, and so it cannot be used to draw conclusions
about why populations differ in a characteristic.
Suppose we measured heritability for human height in
two groups. One group is from a small town in a developed
country, where everyone consumes a high-protein diet.
Because there is little variation in the environmental factors
that affect human height and there is some genetic variation,
the heritability of height in this group is high. The second
group comprises the inhabitants of a single village in a developing country. The consumption of protein by these people
is only 25% of that consumed by those in the first group; so
their average adult height is several centimeters less than that
in the developed country. Again, there is little variation in the
environmental factors that determine height in this group
because everyone in the village eats the same types of food
and is exposed to the same diseases. Because there is little
environmental variation and there is some genetic variation,
the heritability of height in this group also is high.
Thus, the heritability of height in both groups is high, and
the average height in the two groups is considerably different.
We might be tempted to conclude that the difference in height
between the two groups is genetically based—that the people
in the developed country are genetically taller than the people
in the developing country. This conclusion is obviously wrong,
however, because the differences in height are due largely to
diet—an environmental factor. Heritability provides no information about the causes of differences between populations.
These limitations of heritability have often been ignored,
particularly in arguments about possible social implications of
genetic differences between humans. For example, the results
of a number of modern studies indicate that human intelligence as measured by IQ (intelligence quotient) and other
intelligence tests has a moderately high heritability (usually
from 0.4 to 0.8). On the basis of this observation, some people have argued that intelligence is innate and that enhanced
educational opportunities cannot boost intelligence. This
argument is based on the misconception that, when heritability is high, changing the environment will not alter the characteristic. In addition, because heritabilities of intelligence
range from 0.4 to 0.8, a considerable amount of the variance
in intelligence originates from environmental differences.
Another argument based on a misconception about heritability is that ethnic differences in measures of intelligence are
genetically based. Because the results of some genetic studies
show that IQ has moderately high heritability and because
other studies find differences in the average IQ of ethnic
groups, some people have suggested that ethnic differences in
IQ are genetically based. As in the example of the effects of diet
on human height, heritability provides no information about
causes of differences among groups; it indicates only the degree
to which phenotypic variance within a single group is genetically based. High heritability for a characteristic does not mean
that phenotypic differences between ethnic groups are genetic.
We should also remember that separating genetic and environmental effects in humans is very difficult; so heritability estimates themselves may be unreliable.
Locating Genes That Affect
The statistical methods described for use in analyzing quantitative characteristics can be used both to make predictions
about the average phenotype expected in offspring and to
estimate the overall contribution of genes to variation in the
characteristic. These methods do not, however, allow us to
identify and determine the influence of individual genes that
affect quantitative characteristics. As stated in the introduction to this chapter, chromosome regions with genes that
control polygenic characteristics are referred to as quantitative trait loci. Although quantitative genetics has made
important contributions to basic biology and to plant and
animal breeding, the past inability to identify QTLs and
measure their individual effects severely limited the application of quantitative genetic methods.
Mapping QTLs In recent years, numerous genetic markers have been identified and mapped with the use of molecular techniques, making it possible to identify QTLs by linkage
analysis. The underlying idea is simple: if the inheritance of a
genetic marker is associated consistently with the inheritance
of a particular characteristic (such as increased height), then
that marker must be linked to a QTL that affects height. The
key is to have enough genetic markers so that QTLs can be
detected throughout the genome. With the introduction of
restriction fragment length polymorphisms, microsatellite
Heritability provides information only about the degree to which
variation in a characteristic is genetically determined. There is no
universal heritability for a characteristic; heritability is specific for
a given population in a specific environment. Environmental factors can potentially affect characteristics with high heritability,
and heritability says nothing about the nature of population differences in a characteristic.
✔ Concept Check 2
Suppose that you just learned that the narrow-sense heritability of
blood pressure measured among a group of African Americans in
Detroit, Michigan, is 0.40. What does this heritability tell us about
genetic and environmental contributions to blood pressure?
16.13 The availability of molecular markers makes the
mapping of QTLs possible in many organisms. QTL mapping is
used to identify genes that affect yield in corn and other agriculturally
important plants. [Brand X Pictures.]
variations, and single-nucleotide polymorphisms (see Chapter 14), variable markers are now available for mapping QTLs
in a number of different organisms (Figure 16.13; see also the
story at the beginning of this chapter).
A common procedure for mapping QTLs is to cross
two homozygous strains that differ in alleles at many loci.
The resulting F1 progeny are then intercrossed or backcrossed to allow the genes to recombine through independent assortment and crossing over. Genes on different
chromosomes and genes that are far apart on the same
chromosome will recombine freely; genes that are closely
linked will be inherited together. The offspring are measured for one or more quantitative characteristics; at the
same time, they are genotyped for numerous genetic markers that span the genome. Any correlation between the
inheritance of a particular marker allele and a quantitative
phenotype indicates that a QTL is linked to that marker. If
enough markers are used, the detection of all the QTLs
affecting a characteristic is theoretically possible. It is
important to recognize that a QTL is not a gene; rather, a
QTL is a map location for a chromosome region that is
associated with that trait. After a QTL has been identified,
it can be studied for the presence of specific genes that
influence the quantitative trait. This approach has been
used to detect genes affecting various characteristics in several plant and animal species (Table 16.2).
Applications of QTL mapping The number of genes
affecting a quantitative characteristic can be estimated by
locating QTLs with genetic markers and adding up the number of QTLs detected. This method will always be an underestimation because QTLs that are located close together on
the same chromosome will be counted together and those
with small effects are likely to be missed.
QTL mapping also provides information about the
magnitude of the effects that genes have on a quantitative
characteristic. The polygenic model assumes that many
genes affect a quantitative characteristic, that the effect of
each gene is small, and that the effects of the genes are
equal and additive. The results of studies of QTLs in a
number of organisms now show that these assumptions
are not always valid. Polygenes appear to vary widely in
their effects. In many of the characteristics that have been
studied, a few QTLs account for much of the phenotypic
variation. In some instances, individual QTLs have been
mapped that account for more than 20% of the variance in
The availability of numerous genetic markers revealed by molecular methods makes it possible to map chromosome segments containing genes that contribute to polygenic characteristics.
Quantitative characteristics for
which QTLs have been detected
Number of QTLs
Number of ears
Number of nodules
Length of small intestine
Average back fat
Source: After S. D. Tanksley, Mapping polygenes, Annual Review of Genetics
16.4 Genetically Variable
Traits Change in Response
Evolution is genetic change among members of a population. Several different forces are potentially capable of producing evolution, and we will explore these forces and the
process of evolution more fully in Chapter 17. Here, we consider how one of these forces—natural selection—may bring
about genetic change in a quantitative characteristic.
Charles Darwin proposed the idea of natural selection in
his book On the Origin of Species in 1859. Natural selection
arises through the differential reproduction of individuals
with different genotypes, allowing individuals with certain
genotypes to produce more offspring than others. Natural
selection is one of the most important of the forces that
brings about evolutionary change and can be summarized as
Observation 1 Many more individuals are produced
each generation than are capable of surviving long
enough to reproduce.
Observation 2 There is much phenotypic variation
within natural populations.
Observation 3 Some phenotypic variation is heritable.
In the terminology of quantitative genetics, some of
the phenotypic variation in these characteristics is due
to genetic variation, and these characteristics have
Logical consequence Individuals with certain
characters (called adaptive traits) survive and
reproduce better than others. Because the adaptive
traits are heritable, offspring will tend to resemble
their parents with regard to these traits, and there will
be more individuals with these adaptive traits in the
next generation. Thus, adaptive traits will tend to
increase in the population through time.
In this way, organisms become genetically suited to their
environments; as environments change, groups of organisms
change in ways that make them better able to survive and
For thousands of years, humans have practiced a form
of selection by promoting the reproduction of organisms
with traits perceived as desirable. This form of selection is
artificial selection, and it has produced the domestic plants
and animals that make modern agriculture possible.
Predicting the Response to Selection
When a quantitative characteristic is subjected to natural or
artificial selection, it will frequently change with the passage of
time, provided that there is genetic variation for that characteristic in the population. Suppose that a dairy farmer breeds
only those cows in his herd that have the highest milk production. If there is genetic variation in milk production, the mean
milk production in the offspring of the selected cows should
be higher than the mean milk production of the original herd.
This increased production is due to the fact that the selected
cows possess more genes for high milk production than does
the average cow, and these genes are passed on to the offspring.
The offspring of the selected cows possess a higher proportion
of genes for greater milk yield and therefore produce more
milk than the average cow in the initial herd.
The extent to which a characteristic subjected to selection changes in one generation is termed the response to
selection. Suppose that the average cow in a dairy herd produces 80 liters of milk per week. A farmer selects for
increased milk production by breeding the highest milk producers, and the progeny of these selected cows produce 100
liters of milk per week on average. The response to selection
is calculated by subtracting the mean phenotype of the original population (80 liters) from the mean phenotype of the
offspring (100 liters), obtaining a response to selection of
100 Ϫ 80 ϭ 20 liters per week.
The response to selection is determined primarily by
two factors. First, it is affected by the narrow-sense heritability, which largely determines the degree of resemblance
between parents and offspring. When the narrow-sense heritability is high, offspring will tend to resemble their parents;
conversely, when the narrow-sense heritability is low, there
will be little resemblance between parents and offspring.
The second factor that determines the response to selection is how much selection there is. If the farmer is very
stringent in the choice of parents and breeds only the highest milk producers in the herd (say, the top 2 cows), then all
the offspring will receive genes for high-quality milk production. If the farmer is less selective and breeds the top 20 milk
producers in the herd, then the offspring will not carry as
many superior genes for high milk production, and they will
not, on average, produce as much milk as the offspring of the
top 2 producers. The response to selection depends on the
phenotypic difference of the individuals that are selected as
parents; this phenotypic difference is measured by the selection differential, defined as the difference between the mean
phenotype of the selected parents and the mean phenotype
of the original population. If the average milk production of
the original herd is 80 liters and the farmer breeds cows with
an average milk production of 120 liters, then the selection
differential is 120 Ϫ 80 ϭ 40 liters.
The response to selection (R) depends on the narrowsense heritability (h2) and the selection differential (S):
R ϭ h2 ϫ S
This equation can be used to predict the magnitude of
change in a characteristic when a given selection differential
is applied. G. A. Clayton and his colleagues estimated the
response to selection that would take place in abdominal
bristle number of Drosophila melanogaster. Using several
different methods, they first estimated the narrow-sense
heritability of abdominal bristle number in one population
of fruit flies to be 0.52. The mean number of bristles in the
original population was 35.3. They selected individual flies
with a mean bristle number of 40.6 and intercrossed them
to produce the next generation. The selection differential
was 40.6 Ϫ 35.3 ϭ 5.3; so they predicted a response to selection to be
R ϭ 0.52 ϫ 5.3 ϭ 2.8
The response to selection of 2.8 is the expected increase in
the characteristic of the offspring above the mean of the
original population. They therefore expected the average