Linkage and Recombination PDF

har48464_ch05_123-166 07/19/2006 08:52 PM Page 123
CONFIRMING PAGES
PART I Basic Principles: How Traits Are Transmitted
Linkage, Recombination,
and the Mapping of Genes
on Chromosomes
Chapter 5
In 1928, doctors completed a four-generation pedigree
tracing two known X-linked traits: red-green color-
blindness and hemophilia A (the more serious X-linked
form of “bleeders disease”). The maternal grandfather of
the family exhibited both traits, which means that his
single X chromosome carried mutant alleles of the two
corresponding genes. As expected, neither colorblind-
ness nor hemophilia showed up in his sons and daugh-
ters, but two grandsons and one great-grandson inherited
both of the X-linked conditions (Fig. 5.1a). The fact that
none of the descendants of this family’s maternal grand-
father manifested one of the traits without the other
suggests that the mutant alleles did not assort independ-
ently during meiosis. Instead they traveled together in
the gametes forming one generation and then into the
gametes forming the next generation, producing grand-
sons and great-grandsons with an X chromosome speci-
fying both colorblindness and hemophilia. Genes that
travel together more often than not exhibit genetic
linkage. Maps illustrate the spatial
In contrast, another pedigree following colorblindness and the slightly different B relationships of objects, such as
form of hemophilia, which also arises from a mutation on the X chromosome, revealed the locations of subway stations
a different inheritance pattern. A grandfather with hemophilia B and colorblindness along subway lines. Genetic maps
portray the positions of genes
had four grandsons, but only one of them exhibited both conditions. In this family, the
along chromosomes.
genes for colorblindness and hemophilia appeared to assort independently, producing
in the male progeny all four possible combinations of the two traits—normal vision
and normal blood clotting, colorblindness and hemophilia, colorblindness and normal
clotting, and normal vision and hemophilia—in approximately equal frequencies (Fig.
5.1b). Thus, even though the mutant alleles of the two genes were on the same X chro-
mosome in the grandfather, they had to separate to give rise to grandsons III-2 and III-3.
This separation of genes on the same chromosome is the result of recombination, the
occurrence in progeny of new gene combinations not seen in previous generations.
(Note that recombinant progeny can result in either of two ways: from the recombina-
tion of genes on the same chromosome during gamete formation, discussed in this
chapter, or from the independent assortment of genes on nonhomologous chromo-
somes, previously described in Chapter 4.)
As we look at the tools geneticists devised to follow the transmission of genes linked
on the same chromosome, you will see that recombination may separate those genes
when homologous chromosomes exchange parts during meiosis. The farther apart two
genes are, the greater the probability of separation through recombination. Extrapolating
from this general rule, you can see that the gene for hemophilia A must be very close to
the gene for red-green colorblindness, because, as Fig. 5.1a shows, the two rarely sepa-
rate. By comparison, the gene for hemophilia B must lie far away from the colorblindness
gene, because, as Fig. 5.1b indicates, new combinations of alleles of the two genes occur
123
har48464_ch05_123-166 07/19/2006 08:53 PM Page 124
CONFIRMING PAGES
124 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes
(a) I quite often. Geneticists can use data about how often genes sep-
1 2 arate during transmission to map the genes’ relative locations on
II a chromosome. Such mapping is a key to sorting out and track-
1 2
ing down the components of complex genetic networks; it is
III also crucial to geneticists’ ability to isolate and characterize
1 2 3 4 5 6 genes at the molecular level.
IV Three general themes emerge from our discussion of
1 2 Male genetic linkage and recombination. First, maps reflecting
(b) Female the frequency with which linked genes travel together
Hemophilia A helped validate the chromosome theory of inheritance by
I 1 confirming that each chromosome carries many genes in a
Hemophilia B
particular order. Second, the twin concepts of linkage and
II 1 2
Color-blind recombination help explain the genetic underpinnings of
III Hemophilic and evolution. Linkage creates a potential for the simultaneous
1 2 3 4 color-blind transmission of blocks of genes that function well together,
Figure 5.1 Pedigrees indicate that colorblindness and while recombination produces a potential for gene reshuf-
the two forms of hemophilia are X-linked traits. (a) Trans- fling that may enhance the chances of survival under chang-
mission of red-green colorblindness and hemophilia A. Color- ing conditions. Third, chromosome maps led to further
blindness and hemophilia A travel together through the refinement in the concept of a gene. For Mendel, the gene
pedigree, indicating their genetic linkage. (b) Transmission of was an abstract unit that controlled a trait and assorted inde-
red-green colorblindness and hemophilia B. Even though the pendently of other such units. With the chromosome theory,
genes for both conditions are X linked, the mutant alleles are
the gene became part of a chromosome. With mapping, the
inherited together in only one of four grandsons in genera-
gene acquired a chromosomal address—a precise location
tion III. These two pedigrees indicate that the gene for color-
blindness is close to the hemophilia A gene but far away from
that became the focus of more accurate predictions about
the hemophilia B gene on the human X chromosome. inheritance.
Our presentation of the transmission patterns of genes on the same chromosome

describes
• Linkage and meiotic recombination: Genes linked on the same chromosome
usually assort together, instead of independently; such linked genes may
nonetheless become separated through recombination during meiosis.
• Mapping: The frequency with which linked genes become separated through
meiotic recombination generally reflects the physical distance between them;
recombination data thus make it possible to determine the distance between
genes along chromosomes.
• Mitotic recombination: Rarely, recombination occurs during mitosis. In mul-
ticellular organisms, mitotic recombination can produce genetic mosaicism in
which different cells have different genotypes.
genes that do not encode proteins. Recognition that many

genes reside on each chromosome raises an important
5.1 Gene Linkage and question. If genes on different chromosomes assort inde-
Recombination pendently because nonhomologous chromosomes align in-
dependently on the spindle during meiosis I, how do genes
If people have roughly 20,000–30,000 genes but only 23 on the same chromosome assort?
pairs of chromosomes, most human chromosomes must
carry hundreds, if not thousands, of genes. This is certainly
true of the human X chromosome: In 2005, a group of Some Genes on the Same Chromosome
bioinformatics specialists reported that they found 739 pro-
tein-encoding genes on this chromosome. This number is
Assort Together More Often Than Not
likely to grow, at least slightly, as geneticists develop new We begin our analysis with X-linked Drosophila genes be-
techniques to analyze the X chromosome’s DNA sequence. cause they were the first to be assigned to a specific chro-
Moreover, this number does not account for the many mosome and because it is easy to design and follow
har48464_ch05_123-166 07/19/2006 08:53 PM Page 125
CONFIRMING PAGES
5.1 Gene Linkage and Recombination 125
multigenerational crosses yielding large numbers of P

Drosophila progeny. As we outline various crosses, re-
member that females carry two alleles for each X-linked
gene, while males carry only one. This, of course, is be-
cause each female has two X chromosomes, one from each w y+/ w y+ w+ y / Y
parent, whereas each male has only one, the X inherited
F1
from his mother.
We look first at two X-linked genes that determine a
fruit fly’s eye color and body color. These two genes are
said to be syntenic because they are located on the same w y+/ w+ y w y+ / Y
chromosome. The white gene was previously introduced in
Chapter 4; you will recall that the dominant wild-type al-
lele w specifies red eyes, while the recessive mutant allele F2 males
w confers white eyes. The alleles of the yellow body color 4484 w y+ / Y
gene are y (the dominant wild-type allele for brown bod-
ies) and y (the recessive mutant allele for yellow bodies). Parental types =
To avoid confusion, note that lowercase y and y refer to 4484 + 4413 100 ≅ 99%
alleles of the yellow gene, while capital Y refers to the Y 4413 w+ y / Y 9026
chromosome (which does not carry genes for either eye or
body color). You should also pay attention to the slash sym-
bol (/), which is used to separate genes found on chromo-
somes of a pair (either the X and Y chromosomes as in this 76 w+ y+ / Y
case, or a pair of X chromosomes or homologous auto-
somes). Thus w y / Y represents the genotype of a male with Recombinant types =
an X chromosome bearing w and y, as well as a Y chromo- 76 + 53 100 ≅ 1%
9026
some; phenotypically this male has white eyes and a yellow 53 wy/Y
body.
Total 9026
Figure 5.2 When genes are linked, parental combinations
In Dihybrid Crosses, Departures from a 1:1:1:1 outnumber recombinant types. Doubly heterozygous w y /
Ratio of F1 Gametes Indicate That the Two w y F1 females produce four types of sons. Sons that look like
Genes Are on the Same Chromosome the father (w y / Y) or mother (w y / Y) of the F1 females are
In a cross between a female with mutant white eyes and a parental types. Other sons (wy / Y or w y / Y) are recombinant
wild-type brown body (w y/ w y) and a male with wild- types. For these two genes that are very close together on the
Drosophila X chromosome, there are many more parental than
type red eyes and a mutant yellow body (w y / Y), the F1
recombinant types among the progeny.
offspring are evenly divided between brown-bodied
females with normal red eyes (w y / w y) and brown-
bodied males with mutant white eyes (w y / Y) (Fig. 5.2). proximately half of the gametes will be of the two
Note that the male progeny look like their mother because parental types, carrying either the w y allele combina-
their phenotype directly reflects the genotype of the single tion seen in the original mother (the female of the P gen-
X chromosome they received from her. The same is not eration) or the w y allele combination seen in the
true for the F1 females, who received w and y on the original father (the male of the P generation). The re-
X from their mother and w y on the X from their father. maining half of the gametes will be of two recombinant
These F1 females are thus dihybrids: With two alleles for types, in which reshuffling has produced either w y or
each X-linked gene, one derived from each parent, the w y allele combinations not seen in the P generation par-
dominance relations of each pair of alleles determine the ents of the F1 females.
female phenotype. We can see if the 1:1:1:1 ratio of the four kinds of ga-
Now comes the significant cross for answering our metes actually materializes by counting the different
question about the assortment of genes on the same chro- types of male progeny in the F2 generation, as these sons
mosome. If these two Drosophila genes for eye and body receive their only X-linked genes from their maternal ga-
color assort independently, as predicted by Mendel’s sec- mete. The bottom part of Fig. 5.2 depicts the results of a
ond law, the dihybrid F1 females should make four kinds breeding study that produced 9026 F2 males. The relative
of gametes, with four different combinations of genes on numbers of the four X-linked gene combinations passed
the X chromosome—w y, w y, w y, and w y. These on by the dihybrid F1 females’ gametes reflect a signifi-
four types of gametes should occur with equal frequency, cant departure from the 1:1:1:1 ratio expected of inde-
that is, in a ratio of 1:1:1:1. If it happens this way, ap- pendent assortment. By far, the largest numbers of
har48464_ch05_123-166 07/19/2006 08:53 PM Page 126
CONFIRMING PAGES
gametes carry the parental combinations w y and wy. X chromosome genes demonstrates that the observed
Of the total 9026 male flies counted, 8897, or almost 99%, had frequencies of the various types of progeny depend on
these genotypes. In contrast, the new combinations wy how the arrangement of alleles in the F1 females origi-
and w y made up little more than 1% of the total. We can nated. We have redrawn Fig 5.2 as Cross Series A in Fig.
explain why the two genes fail to assort independently in 5.3 so that you can make this comparison more directly.
one of two ways. Either the w y and w y combinations Note that in both experiments, it is the parental classes—
are preferred because of some intrinsic chemical affinity the combinations originally present in the P generation—
between these particular alleles, or it is the parental com- that show up most frequently in the F2 generation. The
bination of alleles the F1 female receives from one or the reshuffled recombinant classes occur less frequently. It
other of her P generation parents that shows up most fre- is important to appreciate that the designation of
quently. “parental” and “recombinant” gametes or progeny of a
doubly heterozygous F1 female is operational, that is, de-
A Preponderance of Parental Genotypes in the termined by the particular set of alleles she receives from
F2 Generation Defines Linkage each of her parents.
When genes assort independently, the numbers of
A second set of crosses involving the same genes but with a parental and recombinant F2 progeny are equal, because a
different arrangement of alleles explains why the dihybrid doubly heterozygous F1 individual produces an equal num-
F1 females do not produce a 1:1:1:1 ratio of the four possi- ber of all four types of gametes. By comparison, two genes
ble types of gametes (see Cross Series B in Fig. 5.3). In this are considered linked when the number of F2 progeny with
second set of crosses, the original parental generation con- parental genotypes exceeds the number of F2 progeny with
sists of red-eyed, brown-bodied females (w y / w y) and recombinant genotypes. Instead of assorting independently,
white-eyed, yellow-bodied males (w y / Y), and the resultant the genes behave as if they are connected to each other
F1 females are all w y/ w y dihybrids. To find out what much of the time. The genes for eye and body color that
kinds and ratios of gametes these F1 females produce, we reside on the X chromosome in Drosophila are an
need to look at the telltale F2 males. extreme illustration of the linkage concept. The two genes
This time, as Cross B in Fig. 5.3 shows, w y / Y and are so tightly coupled that the parental combinations of

w y / Y are the recombinants that account for little more alleles—w y and w y (in Cross Series A of Fig. 5.3) or
than 1% of the total, while w y / Y and w y / Y are the w y and w y (in Cross Series B)—are reshuffled to form
parental combinations, which again add up to almost recombinants in only 1 out of every 100 gametes formed.
99%. You can see that there is no preferred association In other words, the two parental allele combinations of
of w and y or of y and w in this cross. Instead, a these tightly linked genes are inherited together 99 times
comparison of the two experiments with these particular out of 100.
Cross Series A Cross Series B

P P
w y+ w+ y w+ y+ w y
w y+ w+ y+
F1 F1
w y+ w y+ w+ y+ w+ y+
w+ y w y
F2 males F2 males
w y+ w+ y w y w+ y+ w+ y+ w y w+ y w y+
Parental Parental Recombinant Recombinant Parental Parental Recombinant Recombinant

~99% ~1% ~99% ~1%
Figure 5.3 Designations of “parental” and “recombinant” relate to past history. Figure 5.2 has been redrawn here as Cross
Series A for easier comparison with Cross Series B, in which the dihybrid F1 females received a different allelic combination of the white
and yellow genes. Note that the parental and recombinant classes in the two cross series are the opposite of each other. The percent-
ages of recombinant and parental types are nonetheless essentially the same in both experiments, which shows that the frequency of
recombination is independent of the particular arrangement of alleles.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 127
CONFIRMING PAGES
Percentages of Parental and Recombinant P b c +/ b c + b+ c / b+ c

Classes Vary with the Gene Pair
Linkage is not always this tight. In Drosophila, a mutation F1 (all identical) b c+ / b + c
for miniature wings (m) is also found on the X chro-
mosome. A cross of red-eyed females with normal wings Testcross b c + / b+ c b c /b c
(w m/ w m) and white-eyed males with miniature
wings (w m / Y) yields an F1 generation containing all red- Testcross
eyed, normal-winged flies. The genotype of the dihybrid progeny 2934 b c+ / b c Parental 2934 + 2768
classes = 100 = 77%
F1 females is w m/ w m. Of the F2 males, 67.2% are 2768 b+ c / b c 7419
parental types (w m and w m), while the remaining 871 b c / b c Recombinant = 871+ 846
32.8% are recombinants (w m and w m). This prepon- 846 b + c+ /b c classes 7419
100 = 23%
derance of parental combinations among the F2 genotypes Total 7419

reveals that the two genes are linked: The parental combi-
nations of alleles travel together more often than not. But Figure 5.4 Autosomal genes can also exhibit linkage.
compared to the 99% linkage between the w and y genes Genes for body color (b) and wing shape (c) are both autosomal.
for eye color and body color, the linkage of w to m is not A testcross shows that the recombination frequency (RF) for this
that tight. The parental combinations for color and wing pair of genes is 23%. Because parentals outnumber recombi-
size are reshuffled in roughly 33 (instead of 1) out of every nants, the b and c genes are genetically linked and must be on
100 gametes. the same Drosophila autosome.
experiment received parental gene combinations (that

Autosomal Traits Can Also Exhibit Linkage
is, allelic combinations transmitted into the F1 females
Linked autosomal genes are not inherited according to the by the gametes of each of her parents), while the remain-
9:3:3:1 Mendelian ratio expected for two independently as- ing 23% were recombinants. Because the parental
sorting genes. Early twentieth-century geneticists were classes outnumbered the recombinant classes, we can
puzzled by the many experimentally observed departures conclude that the autosomal genes for black body and
from this ratio, which they could not explain in terms of the curved wings are linked.
gene interactions discussed in Chapter 3. They found it To summarize, many pairs of genes on both auto-
difficult to interpret these unexpected results because somes and sex chromosomes exhibit linkage: They do
although they knew that individuals receive two copies of not assort independently; instead they are transmitted to-
each autosomal gene, one from each parent, it was hard to gether more than 50% of the time. (Do not confuse the
trace which alleles came from which parent. However, by general concept of linkage with the idea that some genes
setting up testcrosses in which one parent was homozygous are X linked, which simply means that they reside on the
for the recessive alleles of both genes, they were able to X chromosome.) For each pair of genes, a particular
analyze the gene combinations received from the gametes linkage percentage indicates how often parental combi-
of the other, doubly heterozygous parent. nations travel together. In the preceding examples, the
Fruit flies, for example, carry an autosomal gene for autosomal genes for body color and wing shape were
body color (in addition to the X-linked y gene); the wild transmitted together 77% of the time, the X-chromo-
type is once again brown, but a recessive mutation in this some genes for eye color and wing size were transmitted
gene gives rise to black (b). A second gene on the same together 67.2% of the time, and the X-chromosome
autosome helps determine the shape of a fruit fly’s wing, genes for eye and body color were transmitted together
with the wild type having straight edges and a reces- 99% of the time. Linkage is never 100%. No matter how
sive mutation (c) producing curves. Figure 5.4 depicts a tightly two genes are linked, if you observe enough indi-
cross between black-bodied females with straight wings viduals, you will find some recombinants.
(b c/ b c) and brown-bodied males with curved wings
(b c /b c). All the F1 progeny are double heterozygotes
The Chi-Square Test Pinpoints the Probability
(b c/ b c) that are phenotypically wild type. In a test-
That Experimental Results Are Evidence
cross of the F1 females with b c / b c males, all of the off-
for Linkage
spring receive the recessive b and c alleles from their
father. The phenotypes of the offspring thus indicate the How do you know from a particular experiment whether
kinds of gametes received from the mother. For example, two genes assort independently or are genetically linked?
a black fly with normal wings would be genotype b c/ b At first glance, this question should pose no problem.
c; since we know it received the b c combination from its Discriminating between the two possibilities involves
father, it must have received b c from its mother. As Fig. straightforward calculations based on assumptions well
5.4 shows, roughly 77% of the testcross progeny in one supported by observations. For independently assorting
har48464_ch05_123-166 07/19/2006 08:53 PM Page 128
CONFIRMING PAGES
The Chi-Square Test

The general protocol for using the chi-square test and evaluat- the expected number. Do this for all classes and sum the
ing its results can be stated in a series of steps. Two preparatory individual results (that is, each deviation2 /expected) to-
steps precede the actual chi-square calculation. gether. The final result is the chi-square (2) value. This
step is summarized by the equation
1. Use the data obtained from a breeding experiment to
answer the following questions: 1Number observed Number expected2 2
x2 ©
a. What is the total number of offspring (events) Number expected
analyzed?
b. How many different classes of offspring (events) are where means “sum of all classes.”
there?
c. In each class, what is the number of offspring 4. Compute the degrees of freedom (df). The df is a measure
(events) observed? of the number of independently varying parameters in the
experiment. For the types of experiments we are dis-
2. Calculate how many offspring (events) would be ex- cussing, the number of degrees of freedom is one less than
pected for each class if the null hypothesis (here, no link- the number of classes. For example, if the total number of
age) were correct by multiplying the percentage offspring in a testcross fall into four classes and you know
predicted by the null hypothesis (here, 50% parentals the number present in any three, you can easily calculate
and 50% recombinants) by the total number of off- the number in the fourth. Thus, if N the number of
spring analyzed in the experiment. You are now ready classes, then the degrees of freedom (df) N 1. If there
for the chi-square calculation. are 4 classes, there are only 3 df.
3. To calculate chi square, begin with one class of offspring. 5. Use the chi-square value together with the number of
Subtract the expected number from the observed num- degrees of freedom to determine a p value: the proba-
ber (which gives the deviation from the prediction for bility that a deviation from the predicted numbers at
this class), square the result, and divide this amount by least as large as that observed in the experiment will
genes, a dihybrid F1 female produces four types of gametes particular hypothesis will occur solely by chance. This
in equal numbers, so one-half of the F2 progeny are of the measure of the “goodness of fit” between observed and pre-
parental classes and the other half of the recombinant dicted results is a probability test known as the chi-square
classes. In contrast, for linked genes, the two types of test. The test is designed to account for the fact that the size
parental classes by definition always outnumber the two of an experimental population (the “sample size”) is an im-
types of recombinant classes in the F2 generation. portant component of statistical significance. To appreciate
The problem is that because real-world genetic trans- the role of sample size, it is useful to return to the proverbial
mission is based on chance events, in a particular study coin toss before examining the details of the chi-square test.
even unlinked, independently assorting genes can produce We have seen that in 10 tosses of a coin, because of
deviations from the 1:1:1:1 ratio, just as in 10 tosses of a chance, an outcome of 6 heads (60%) and 4 tails (40%) is
coin, you may easily get 6 heads and 4 tails (rather than the not unexpected. In contrast, with 1000 tosses of the coin, a
predicted 5 and 5). Thus, if a breeding experiment analyz- result of 600 heads (60%) and 400 tails (40%) would intu-
ing the transmission of two genes shows a deviation from itively be highly unlikely. In the first case, a change in the
the equal ratios of parentals and recombinants expected of results of one coin toss would alter the expected 5 heads
independent assortment, can we necessarily conclude the and 5 tails to the observed 6 heads and 4 tails. In the second
two genes are linked? Is it instead possible that the results case, 100 tosses would have to change from tails to heads
represent a statistically acceptable chance fluctuation from to generate the stated deviation from the predicted 500
the mean values expected of unlinked genes that assort in- heads and 500 tails. It is reasonable, even likely, that
dependently? Such questions become more pressing in chance events could cause 1 deviation from the predicted
cases where linkage is not all that tight, so that even though number, but not 100. Two important concepts emerge from
the genes are linked, the percentage of recombinant classes this simple presentation. First, a comparison of percentages
approaches 50%. or ratios alone will never allow you to determine whether
To answer these kinds of questions, statisticians have or not observed data are significantly different from pre-
devised a quantitative measure that indicates how often an dicted values. Second, the absolute numbers obtained are
experimentally observed deviation from the predictions of a important because they reflect the size of the experiment.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 129
CONFIRMING PAGES
occur by chance. Although the p value is arrived at same size, the observed data showing a deviation from
through a numerical analysis, geneticists routinely deter- the predictions of the null hypothesis could have been
mine the value by a quick search through a table of crit- obtained by chance even if the null hypothesis were
ical 2 values, such as the one in Table 5.1. true; the data are therefore not significant for rejecting
the null hypothesis and showing linkage. Statisticians
6. Evaluate the significance of the p value. You can think have arbitrarily selected the 0.05 p value as the bound-
of the p value as the probability that the null hypothe- ary between accepting and rejecting the null hypothe-
sis is true and the alternative hypothesis (in this case, sis. A p value of less than 0.05 means that you can
that the genes are linked) is wrong. A value greater consider the data showing deviation significant and re-
than 0.05 indicates that in more than 1 in 20 (or more ject the null hypothesis.
than 5%) of the repetitions of an experiment of the
TABLE 5.1 Critical Chi-Square Values

p Values
Cannot Reject the Null Hypothesis Null Hypothesis Rejected

Degrees of
Freedom 0.99 0.90 0.50 0.10 0.05 0.01 0.001
2
Values
1 — 0.02 .45 2.71 3.84 6.64 10.83
2 0.02 0.21 1.39 4.61 5.99 9.21 13.82
3 0.11 0.58 2.37 6.25 7.81 11.35 16.27

4 0.30 1.06 3.36 7.78 9.49 13.28 18.47
5 0.55 1.61 4.35 9.24 11.07 15.09 20.52
Note: 2 values that lie in the reddish-shaded region of this table allow you to reject the null hypothesis with 95% confidence, and for recombina-
tion experiments, to postulate linkage.
The larger the sample size, the closer the observed percent- that is, they differ enough not to be reasonably attributable
ages can be expected to match the values predicted by the to chance alone, then the researcher can reject the null hy-
experimental hypothesis, if the hypothesis is correct. The pothesis of no linkage and accept the alternative of linkage
chi-square test is therefore always calculated with num- between the two genes.
bers—actual data—and not percentages or proportions. The Tools of Genetics box beginning on p. 128 pres-
A critical prerequisite of the chi-square test is the ents the general protocol of the chi-square test. The final re-
framing of a hypothesis that leads to clear-cut predictions. sult of the calculations is the determination of the
Although contemporary geneticists use the chi-square test numerical probability—the p value—that a particular set of
to interpret many kinds of genetic experiments, they use it observed experimental results represents a chance devia-
most often to discover whether data obtained from breed- tion from the values predicted by a particular hypothesis. If
ing experiments provide evidence for or against the hy- the probability is high, it is likely that the hypothesis being
pothesis that two genes are linked. But the problem with tested explains the data, and the observed deviation from
the general hypothesis that “genes A and B are linked” is expected results is considered insignificant. If the probabil-
that there is no precise prediction of what to expect in terms ity is very low, the observed deviation from expected re-
of breeding data. This is because, as we have seen, the fre- sults becomes significant. When this happens, it is unlikely
quency of recombination varies with each linked gene pair. that the hypothesis under consideration explains the data,
In contrast, the alternative hypothesis “that genes A and B and the hypothesis can be rejected.
are not linked” gives rise to a precise prediction: that alle-
les at different genes will assort independently and produce
Applying the Chi-Square Test
50% parental and 50% recombinant progeny. So, whenever
a geneticist wants to determine whether two genes are Figure 5.5 depicts two sets of data obtained from testcross
linked, he or she actually tests whether the observed data experiments asking whether genes A and B are linked. We
are consistent with a null hypothesis of no linkage. If the first apply the chi-square analysis to data accumulated in
chi-square test shows that the observed data differ signifi- the first experiment. The total number of offspring is 50, of
cantly from those expected with independent assortment, which 31 (that is, 17 14) are observed to be parental
har48464_ch05_123-166 07/19/2006 08:53 PM Page 130
CONFIRMING PAGES
Progeny Experiment 1 Experiment 2 A certain amount of subjectivity enters into any deci-
sion about the significance of a particular p value. As
AB 17 34
stated in the preceding box, statisticians have arbitrarily
ab 14 28
Ab 8 16 selected a p value of 0.05 as the boundary between signifi-
aB 11 22 cance and nonsignificance. Values lower than this indicate
there would be less than 5 chances in 100 of obtaining
Total 50 100 the same results by random sampling if the null hypothesis
Class Observed / Expected Observed / Expected were true. An extremely low p value thus suggests that
Parentals 31 25 62 50 the data shows major deviations from values predicted
Recombinants 19 25 38 50 by the hypothesis; these values are significant enough to
reject the null hypothesis. The lower the p value, the clearer
Figure 5.5 Applying the chi-square test to see if genes A the case that the data cannot be regarded as chance devia-
and B are linked. The null hypothesis is that the two genes are tions from the predictions of the hypothesis. For this
unlinked. For Experiment 1, p 0.05, so it is not possible to reject reason, more conservative scientists often set the boundary
the null hypothesis. For Experiment 2, with a data set twice the of significance at p 0.01 and would thus be less willing
size, p 0.05. This is below the arbitrary boundary of significance to reach a conclusion based on the results of Experiment 2
used by most geneticists, which makes it possible to reject the in Fig. 5.5. In linkage studies, such a low p value means
null hypothesis and conclude with greater than 95% confidence that it is highly unlikely (only 1 chance in 100) that the ob-
that the genes are linked. served distribution of phenotypes arose from two unlinked
genes, and we can conclude, with 99% confidence, that the
genes are linked. On the other hand, higher p values
types and 19 (8 11) recombinant types. Dividing 50 by 2, (greater than 0.01 or 0.05, depending on the criterion used)
you get 25, the number of parental or recombinant do not necessarily mean that two genes are unlinked; it may
offspring expected according to the null hypothesis of mean only that the sample size is not large enough to
independent assortment (which predicts that parentals provide an answer. With more data, the p value will
recombinants). Now, considering first the parental types normally rise if the null hypothesis of no linkage is correct
alone, you square the observed deviation from the expected and fall if there is, in fact, linkage.
value, and divide the result by the expected value. After do- Note that in Fig. 5.5 all of the numbers in the second
ing the same for the recombinant types, you add the two set of data are simply double the numbers in the first set,
quotients to obtain the value of chi square. with the percentages remaining the same. Thus, just by
doubling the sample size from 50 to 100 individuals, it was
131 252 2 119 252 2 possible to go from no significant difference to a significant
x2 1.44 1.44 2.88
25 25 difference between the observed and the expected values
(using the p 0.05 cutoff). In other words, the larger the
Since with two classes (parentals and recombinants), the sample size, the less the likelihood that a certain percentage
number of degrees of freedom is 1, you scan the chi-square deviation from expected results happened simply by
table (see Table 5.1 on p. 129) for 2 2.88 and df 1, and chance. Bearing this in mind, you can see that it is not ap-
you find by extrapolation that the corresponding p value is propriate to use the chi-square test when analyzing very
greater than 0.05 (roughly 0.09). From this, you can con- small samples of less than 10, which by their nature are
clude that it is not possible to reject the null hypothesis on usually insufficient to answer questions concerning link-
the basis of this experiment, which means that this data set age. This creates a problem for human geneticists, who
is not sufficient to demonstrate linkage between A and B. cannot “construct” people as they can fruit flies and other
If you use the same strategy to calculate a p value for experimental organisms nor develop large populations of
the data observed in the second experiment, where there inbred individuals. To achieve a reasonable sample size for
are a total of 100 offspring and thus an expected number of linkage studies in humans, they must pool data from a large
50 parentals and 50 recombinants, you get number of family pedigrees.
162 502 2 138 502 2 The subjectivity involved in determining the bound-
x2 2.88 2.88 5.76 ary of significance means that the chi-square test does not
50 50
prove linkage or its absence. What it does do is provide a
The number of degrees of freedom (df ) remains 1, so Table quantitative measure of the likelihood that the data from
5.1 arrives at a p value greater than 0.01 but less than 0.05. an experiment can be explained by a particular hypothe-
In this case, you can consider the difference between the sis. The chi-square analysis is thus a general statistical
observed and expected values to be significant. As a result, test for significance; it can be used with many different
you can reject the null hypothesis of independent assort- experimental designs and with hypotheses other than the
ment and conclude it is likely that genes A and B are linked. absence of linkage. As long as it is possible to propose a
har48464_ch05_123-166 07/19/2006 08:53 PM Page 131
CONFIRMING PAGES
hypothesis that leads to a predicted set of values for a de- Reciprocal Exchanges Between
fined set of data classes, you can readily determine Homologous Chromosomes Are the Physical
whether or not the observed data are consistent with the Basis of Recombination
hypothesis. If what you are looking for is any one of a
range of outcomes that is not the outcome predicted by Morgan’s idea that the physical breaking and rejoining of
the hypothesis, that is, if the hypothesis serves as a “straw chromosomes during meiosis was the basis of genetic
man,” set up only to be knocked down, it is considered a recombination seemed reasonable. But although Janssens’s
null hypothesis. chiasmata could be interpreted as signs of the process,
When experiments lead to rejection of a null hypothe- before 1930 no one had produced visible evidence that
sis, you may need to confirm the alternative. For instance, crossing-over between homologous chromosomes actually
if you are analyzing the inheritance of two opposing traits occurs. The identification of physical markers, or cytologi-
on the basis of a null hypothesis that says the two pheno- cally visible abnormalities that make it possible to keep track
types result from the segregation of two equally viable al- of specific chromosome parts from one generation to the
leles of a single gene, you would expect a testcross next, enabled researchers to turn the logical deductions about
between an F1 heterozygote and a recessive homozygote to recombination into facts derived from experimental evidence.
produce a 1:1 ratio of the two traits in the offspring. If in- In 1931, Harriet Creighton and Barbara McClintock, who
stead, you observe a ratio of 6:4 and the chi-square test studied corn, and Curt Stern, who worked with Drosophila,
produces a p value of 0.009, you can reject the null hy- published the results of experiments showing that genetic re-
pothesis. But you are still left with the question of what the combination indeed depends on the reciprocal exchange of
absence of a 1:1 ratio means. There are actually two alter- parts between maternal and paternal chromosomes. Stern,
natives: Either the reverse of the null hypothesis is true and for example, bred female flies with two different X chromo-
the two alleles of the single gene are not equally viable, or somes, each containing a distinct physical marker near one
more than one gene encodes the trait. The chi-square test of the ends. These same females were also doubly heterozy-
cannot tell you which possibility is correct, and you would gous for two X-linked genetic markers—genes that could
have to study the matter further. The problems at the end of serve as points of reference in determining whether particu-
this chapter illustrate several applications of the chi-square lar progeny were the result of recombination.
test pertinent to genetics. Figure 5.6 diagrams the chromosomes of these het-
erozygous females. One X chromosome carried mutations
producing carnation eyes (a dark ruby color, abbreviated
car) that were kidney-shaped (Bar); in addition, this chro-
Recombination Results When mosome was marked physically by a visible discontinuity,
Crossing-Over During Meiosis which resulted when the end of the X chromosome was
broken off and attached to an autosome. The other X chro-
Separates Linked Genes mosome had wild-type alleles () for both the car and
It is easy to understand how genes that are physically con- the Bar genes, and its physical marker consisted of part of
nected on the same chromosome can be transmitted to- the Y chromosome that had become connected to the
gether and thus show genetic linkage. It is not as obvious X-chromosome centromere.
why all linked genes always show some recombination in a Figure 5.6 illustrates how the chromosomes in these car
sample population of sufficient size. Do the chromosomes Bar / car Bar females were transmitted to male progeny.
participate in a physical process that gives rise to the According to the experimental results, all sons showing a
reshuffling of linked genes that we call recombination? The phenotype determined by one or the other parental combina-
answer to this question is of more than passing interest as it tion of genes (either car Bar or car Bar) had an X chro-
provides a basis for gauging relative distances between mosome that was structurally indistinguishable from one
pairs of genes on a chromosome. of the original X chromosomes in the mother. In recombi-
In 1909, the Belgian cytologist Frans Janssens de- nant sons, however, such as those that manifested carnation
scribed structures he had observed in the light microscope eye color and normal eye shape (car Bar/Y), an identifi-
during prophase of the first meiotic division. He called able exchange of the abnormal features marking the ends
these structures chiasmata; as described in Chapter 4, they of the homologous X chromosomes accompanied the re-
seemed to represent regions in which nonsister chromatids combination of genes. The evidence thus tied a particular
of homologous chromosomes cross over each other (review instance of phenotypic recombination to the crossing-over
Fig. 4.14 on p. 98). Making inferences from a combination of particular genes located in specifically marked parts of
of genetic and cytological data, Thomas Hunt Morgan sug- particular chromosomes. This was an elegant demonstra-
gested that the chiasmata observed through the light micro- tion that genetic recombination is associated with the ac-
scope were sites of chromosome breakage and exchange tual reciprocal exchange of segments between homologous
resulting in genetic recombination. chromosomes during meiosis.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 132
CONFIRMING PAGES
Additional material from

part of the Y chromosome Discontinuity
car Bar
Parental ( )
chromosomes
car+ Bar+
Meiosis
No crossing-over Crossing-over
car Bar car Bar
car+ Bar+ car+ Bar+
car Bar car Bar+
Chromosomes Parental Recombinant

transmitted to
progeny ( ) car+ Bar+ car+ Bar
Parental Recombinant
Figure 5.6 Evidence that recombination results from reciprocal exchanges between homologous chromosomes. Genetic
recombination between the car and Bar genes on the Drosophila X chromosome is accompanied by the exchange of physical markers
observable in the microscope. Note that this depiction of crossing-over is a simplification, as genetic recombination actually occurs after
each chromosome has replicated into sister chromatids. Note also that the piece of the X chromosome to the right of the discontinuity is
actually attached to an autosome.
Through the Light Microscope: Chiasmata Mark Visible in the light microscope, chiasmata indicate where
the Sites of Recombination chromatid sections have switched from one molecule to an-
other. In Fig. 5.7d, during anaphase I, as the two homologs
Figure 5.7 outlines what is currently known about the steps separate, starting at their centromeres, the ends of the two re-
of recombination as they appear in chromosomes viewed combined chromatids pull free of their respective sister chro-
through the light microscope. Although this low-resolution matids, and the chiasmata shift from their original positions
view may not represent certain details of recombination with toward a chromosome end, or telomere. This movement of
complete accuracy, it nonetheless provides a useful frame of chiasmata is known as terminalization. When the chiasmata
reference. In Fig. 5.7a, the two homologs of each chromo- reach the telomeres, the homologous chromosomes can sep-
some pair have already replicated, so there are now two pairs arate from each other (Fig. 5.7e). Meiosis continues and
of sister chromatids or a total of four chromatids within each eventually produces four haploid cells that contain one
bivalent. In Fig. 5.7b, the synaptonemal complex zips to- chromatid—now a chromosome—apiece (Fig. 5.7f ). Ho-
gether homologous chromosome pairs along their length. mologous chromosomes have exchanged parts.
The synaptonemal zipper aligns homologous regions of all Recombination can also take place apart from meiosis.
four chromatids such that allelic DNA sequences are physi- Indeed, as we explain near the end of this chapter, it some-
cally near each other (see Fig. 4.14b on p. 98 for a detailed times, though rarely, occurs during mitosis. It also occurs
depiction). This proximity facilitates crossing-over between with the circular chromosomes of prokaryotic organisms
homologous sequences; as we will see in Chapter 6, the bio- and with cellular organelles such as mitochondria and
chemical mechanism of recombination requires a close in- chloroplasts, which do not undergo meiosis and do not
teraction of DNAs on homologous chromosomes that have form chiasmata (see Chapters 15 and 16).
identical, or nearly identical, nucleotide sequences.
In Fig. 5.7c, the synaptonemal complex begins to disas-
Recombination Frequencies for Pairs of Genes
semble. Although at least some steps of the recombination
Reflect the Distances Between Them
process occurred while the chromatids were zipped in synap-
sis, it is only now that the recombination event becomes ap- Thomas Hunt Morgan’s belief that chiasmata represent
parent. As the zipper dissolves, homologous chromosomes sites of physical crossing-over between chromosomes and
remain attached at chiasmata, the actual sites of crossing-over. that such crossing-over may result in recombination, led
har48464_ch05_123-166 07/19/2006 08:53 PM Page 133
CONFIRMING PAGES
(a) Duplicated chromosome homologs (a) w y
1 1.1 m.u.
2 (b)
w m
3 32.8 m.u.
4 Figure 5.8 Recombination frequencies are the basis of

genetic maps. (a) 1.1% of the gametes produced by a female
doubly heterozygous for the genes w and y are recombinant.
(b) Synapsis
The recombination frequency (RF) is thus 1.1%, and the genes are
approximately 1.1 map units (m.u.) or 1.1 centimorgans (cM)
apart. (b) The distance between the w and m genes is longer:
32.8 m.u. (or 32.8 cM).
Chiasmata become visible
(sites of crossing-over)
(c)
him to the following logical deduction: Different gene
pairs exhibit different linkage rates because genes are
arranged in a line along a chromosome. The closer together
two genes are on the chromosome, the less their chance of
being separated by an event that cuts and recombines the
(d) Terminalization
line of genes. To look at it another way, if we assume for
the moment that chiasmata can form anywhere along a
chromosome with equal likelihood, then the probability of
a crossover occurring between two genes increases with
the distance separating them. If this is so, the frequency of
genetic recombination also must increase with the distance
between genes. To illustrate the point, imagine pinning to a
Anaphase I
wall 10 inches of ribbon with a line of tiny black dots along
(e)
Segregation of its length and then repeatedly throwing a dart to see where
homologous chromosomes
you will cut the ribbon. You would find that practically
every throw of the dart separates a dot at one end of the rib-
bon from a dot at the other end, while few if any throws
separate any two particular dots positioned right next to
each other.
Alfred H. Sturtevant, one of Morgan’s students, took
this idea one step further. He proposed that the percentage
(f) Meiosis II of total progeny that were recombinant types, the recombi-
Haploid products nation frequency (RF), could be used as a gauge of the
physical distance separating any two genes on the same
chromosome. Sturtevant arbitrarily defined one RF per-
centage point as the unit of measure along a chromosome;
later, another geneticist named the unit a centimorgan
(cM) after T. H. Morgan. Mappers often refer to a centi-
morgan as a map unit (m.u.). Although the two terms are
interchangeable, researchers prefer one or the other,
depending on their experimental organism. Drosophila
Figure 5.7 Recombination through the light microscope. geneticists, for example, use map units while human
(a) A pair of duplicated homologous chromosomes very early in geneticists use centimorgans. In Sturtevant’s system, 1%
prophase of meiosis I. (b) During leptotene and zygotene of
RF 1 cM 1 m.u. A review of the two pairs of X-linked
prophase I, the synaptonemal complex helps align corresponding
Drosophila genes we analyzed earlier shows how his pro-
regions of homologous chromosomes, allowing recombination.
(c) As the synaptonemal complex disassembles during diplotene,
posal works. Because the X-linked genes for eye color (w)
homologous chromosomes remain attached at chiasmata. (d) and and body color (y) recombine in 1.1% of F2 progeny, they
(e) The chiasmata terminalize (move toward the chromosome are 1.1 m.u. apart (Fig. 5.8a). In contrast, the X-linked
ends), allowing the recombined chromosomes to separate during genes for eye color (w) and wing size (m) have a recombi-
anaphase and telophase. (f) The result of the process is recombination frequency of 32.8 and are therefore 32.8 m.u. apart
nant gametes. (Fig. 5.8b).
har48464_ch05_123-166 07/19/2006 08:53 PM Page 134
CONFIRMING PAGES
As a unit of measure, the map unit is simply an index classes are equal), you can demonstrate they are on the
of recombination probabilities assumed to reflect distances same chromosome if you can tie each of the widely sepa-
between genes. According to this index, the y and w genes rated genes to one or more common intermediaries.
are much closer together than the m and w genes. Geneti-
cists have used this logic to map thousands of genetic
markers to the chromosomes of Drosophila, building re-
combination maps step-by-step with closely linked mark-
Linkage and Recombination: A Summary
ers. And as we see next, they have learned that markers The important conclusions from the experimental data pre-
very far apart on the same chromosome may appear un- sented thus far in this chapter can be summarized as fol-
linked, even though their recombination distances relative lows. Gene pairs that are close together on the same
to closely linked intervening markers confirm that they are chromosome are linked and do not follow Mendel’s law of
indeed on the same chromosome. independent assortment. Instead, alleles of linked genes are
more or less tightly coupled during transmission, leading to
a preponderance of parental classes among the progeny of
Experimental Recombination Frequencies
double heterozygotes. Linked alleles, however, can be-
Between Two Genes Are Never
come separated through recombination, and the frequency
Greater Than 50%
with which this happens is different for each pair of genes.
If the definition of linkage is that the proportion of recom- There is a correlation between recombination frequency
binant classes is less than that of parental classes, a recom- and the distance separating two genes.
bination frequency of less than 50% indicates linkage. But The mechanism of recombination is crossing-over:
what can we conclude about the relative location of genes Some genes move from the maternal to the paternal chro-
if there are roughly equal numbers of parental and recom- mosome, and vice versa, when homologs exchange parts
binant progeny? And does it ever happen that recombinants during meiosis. Chiasmata are the visible signs of crossing-
are in the majority? over, and if we assume that chiasmata can occur anywhere
We already know one situation that can give rise to a re- along a chromosome, the farther apart two genes are, the
combination frequency of 50%. Genes located on different greater the opportunity for chiasmata to form between
(that is, nonhomologous) chromosomes will obey Mendel’s them. This explains why larger recombination frequencies
law of independent assortment because the two chromo- reflect greater distances between genes. Recombination
somes can line up on the spindle during meiosis I in either of frequencies in pairwise crosses vary from 0% to 50%. Sta-
two equally likely configurations (review Fig. 4.17a on p. tistically significant values of less than 50% indicate that
101). A dihybrid for these two genes will thus produce all two genes are linked and must therefore be on the same
four possible types of gametes (AB, Ab, aB, and ab) with ap- chromosome. Two genes that show a recombination fre-
proximately equal frequency. Importantly, experiments have quency of 50% are genetically unlinked, either because
established that genes located very far apart on the same chro- they are on different chromosomes or because they reside
mosome also show recombination frequencies of approxi- far apart on the same chromosome (Table 5.2).
mately 50%. Knowledge of linkage and recombination paves the
Researchers have never observed statistically signifi- way for understanding how geneticists assign genes a rel-
cant recombination frequencies between two genes greater ative chromosomal position by comparing the recombina-
than 50%, which means that in any cross following two tion frequencies of many gene pairs. The chromosomal
genes, recombinant types are never in the majority. As we position assigned to a gene is its locus. Each gene’s locus
explain in more detail later in the chapter, this upper limit
of 50% on the recombination frequency between two genes
results from two aspects of chromosome behavior during
meiosis I. First, multiple crossovers can occur between two TABLE 5.2 Properties of Linked Versus
genes if they are far apart on the same chromosome, and Unlinked Genes
second, recombination takes place after the chromosomes
have replicated into sister chromatids. Linked Genes
For now, simply note that recombination frequencies Parentals recombinants (RF 50%)
near 50% suggest either that two genes are on different Linked genes must be syntenic and sufficiently close
together on the same chromosome so that they do not assort
chromosomes or that they lie far apart on the same chro-
independently.
mosome. The only way to tell whether the two genes are
syntenic (that is, on the same chromosome) is through a se- Unlinked Genes
ries of matings showing definite linkage with other genes Parentals recombinants (RF 50%)
that lie between them. In short, even though crosses be- Occurs either when genes are on different chromosomes or
tween two genes lying very far apart on a chromosome may when they are sufficiently far apart on the same chromosome.
show no linkage at all (because recombinant and parental
har48464_ch05_123-166 07/19/2006 08:53 PM Page 135
CONFIRMING PAGES
5.2 Mapping; Locating Genes Along a Chromosome 135
is the same in all individuals of a species, and the process (a) Gene (b) y w m
pair RF
of determining that locus is known as mapping. As you
will see, the results of mapping experiments consistently y-w 1.1 1.1 32.8
verify the idea that genes are arranged in a line along a y -v 33.0 34.3
y-m 34.3 (c) y w v m r
chromosome. y -r 42.9
w -v 32.1
w-m 32.8 1.1
w -r 42.1 33.0
v-m 4.0 34.3
v -r 24.1 42.9
5.2 Mapping: Locating Genes m -r 17.8 (d) y w v m r
Along a Chromosome 1.1+ 32.1 + 4.0 + 17.8 = 5
Figure 5.9 Mapping genes by comparisons of two-point

Maps are images of the relative positions of objects in crosses. (a) Sturtevant’s data for the distances between pairs of
space. Whether depicting the floor plan of New York’s Met- X-linked genes in Drosophila. (b) Because the distance between y
ropolitan Museum of Art, the layout of the Roman Forum, and m is greater than the distance between w and m, the order
or the location of cities served by the railways of Europe, of genes must be y-w-m. (c) and (d) Maps for five genes on the
maps turn measurements (how far apart two rooms, ruins, or Drosophila X chromosome. The left-to-right orientation is
railway stops are from one another) into patterns of spatial arbitrary. Note that the numerical position of the r gene depends
relationships that add a new level of meaning to the original on how it is calculated. The best genetic maps are obtained by
summing many small intervening distances as in (d).
data of distances. Maps that assign genes to locations on
particular chromosomes are no exception. By transforming
genetic data into spatial arrangements, they sharpen our
ability to predict the inheritance patterns of specific traits. Recombination Mapping Supports the Idea
We have seen that recombination frequency (RF) is a That Genes Are Arranged in a Line Along
measure of the distance separating two genes along a chro- a Chromosome
mosome. We now examine how data from many crosses
By following exactly the same procedure for each set of
following two and three genes at a time can be compiled
three genes, Sturtevant established a self-consistent order for
and compared to generate accurate, comprehensive gene/
all the genes he investigated on Drosophila’s X chromosome
chromosome maps.
(Fig. 5.9c; once again, the left-to-right arrangement is an
arbitrary choice). By checking the data for every combina-
tion of three genes, you can assure yourself that this ordering
Two-Point Crosses: Comparisons Help makes sense. The fact that the recombination data yield a
simple linear map of gene position supports the idea that
Establish Relative Gene Positions genes reside in a unique linear order along a chromosome.
In his senior undergraduate thesis, Morgan’s student A. H.
Sturtevant asked whether data obtained from a large number
Two-Point Crosses Have Their Limitations
of two-point crosses (crosses tracing two genes at a time)
would support the idea that genes form a definite linear se- Though of great importance, the pairwise mapping of
ries along a chromosome. He began by looking at X-linked genes has several shortcomings that limit its usefulness.
genes in Drosophila. Figure 5.9a lists his recombination First, in crosses involving only two genes at a time, it may
data for several two-point crosses. Recall that the distance be difficult to determine gene order if some gene pairs lie
between two genes that yields 1% recombinant progeny— very close together. For example, in mapping y, w, and m,
an RF of 1%—is 1 m.u. As an example of Sturtevant’s rea- 34.3 m.u. separate the outside genes y and m, while nearly
soning, consider the three genes w, y, and m. If these genes as great a distance (32.8 m.u.) separates the middle w from
are arranged in a line (instead of a more complicated the outside m (Fig. 5.9b). Before being able to conclude with
branched structure, for example), then one of them must be any confidence that y and m are truly farther apart, that is,
in the middle, flanked on either side by the other two. The that the small difference between the values of 34.3 and 32.8
greatest genetic distance should separate the two genes on is not the result of sampling error, you would have to exam-
the outside, and this value should roughly equal the sum of ine a very large number of flies and subject the data to a sta-
the distances separating the middle gene from each outside tistical test, such as the chi-square test.
gene. The data Sturtevant obtained are consistent with this A second problem with Sturtevant’s mapping procedure
idea, implying that w lies between y and m (Fig. 5.9b). Note is that the actual distances in his map do not always add up,
that the left-to-right orientation of this map was selected at even approximately. As an example, suppose that the locus
random; the map in Fig. 5.9b would be equally correct if it of the y gene at the far left of the map is regarded as position
portrayed y on the right and m on the left. 0 (Fig. 5.9c). The w gene would then lie near position 1, and
har48464_ch05_123-166 07/19/2006 08:53 PM Page 136
CONFIRMING PAGES
m would be located in the vicinity of 34 m.u. But what about

(a) Three-point cross results
the r gene, named for a mutation that produces rudimentary
(very small) wings? Based solely on its distance from y, as in- P vg b pr / vg b pr vg+ b+ pr + / vg+ b+ pr +
ferred from the y ↔ r data in Fig. 5.9a, we would place it at
position 42.9 (Fig. 5.9c). However, if we calculate its position F1 (all identical) vg b pr / vg+ b+ pr +
as the sum of all intervening distances inferred from the data
in Fig. 5.9a, that is, as the sum of y ↔ w plus w ↔ v plus v ↔
m plus m ↔ r, the locus of r becomes 1.1 32.1 4.0 Testcross vg b pr / vg+ b+ pr + vg b pr / vg b pr
17.8 55.0 (Fig. 5.9d). What can explain this difference, and
which of these two values is closer to the truth? Three-point
crosses help provide some of the answers. Testcross 1779 vg b pr Parental combinations for
progeny 1654 vg+ b+ pr + all three genes
252 vg+ b pr Recombinants for vg relative to
vg b+ pr +
Three-Point Crosses: A Faster, More 241
131 vg+ b pr +
parental combinations for b and pr
Recombinants for b relative to
Accurate Way to Map Genes 118
13
vg b+ pr
vg b pr +
parental combinations for vg and pr
Recombinants for pr relative to
The simultaneous analysis of three markers makes it possi- 9 vg+ b+ pr parental combinations for vg and b
ble to obtain enough information to position the three 4197

genes in relation to each other from just one set of crosses. (b) Deduced genetic map
To describe this procedure, we look at three genes linked vg pr b
on one of Drosophila’s autosomes.
A homozygous female with mutations for vestigial
12.3 m.u. 6.4 m.u. = 18.7 m.u.
wings (vg), black body (b), and purple eye color (pr) was
mated to a wild-type male (Fig. 5.10a). All the triply het- 17.7 m.u.
erozygous F1 progeny, both male and female, had normal Figure 5.10 Analyzing the results of a three-point cross.
phenotypes for the three characteristics, indicating that the (a) Results from a three-point testcross of F1 females simultane-
mutations are autosomal recessive. In a testcross of the F1 ously heterozygous for vg, b, and pr. (b) The gene in the middle
females with males having vestigial wings, black body, and must be pr because the longest distance is between the other
purple eyes, the progeny were of eight different phenotypes two genes: vg and b. The most accurate map distances are calcu-
lated by summing shorter intervening distances, so 18.7 m.u. is a
reflecting eight different genotypes. The order in which the
more accurate estimate of the genetic distance between vg and b
genes in each phenotypic class are listed in Fig. 5.10a is
than 17.7 m.u.
completely arbitrary; instead of vg b pr, one could write b
vg pr or vg pr b to indicate the same genotype. This is be-
cause at the outset, we do not know the gene order. Deduc-
while the distance separating the b–pr pair is
ing it is the goal of the mapping study. Thus, the way the
data are tabulated does not necessarily indicate the actual 131 118 13 9
order of genes along the chromosome. 100
4197
In analyzing the data, we look at two genes at a time (re-
call that the recombination frequency is always a function of 6.4 m.u. 1b 4 pr distance2
a pair of genes). For the pair vg and b, the parental combina- These recombination frequencies show that vg and b
tions are vg b and vgb; the nonparental recombinants are are separated by the largest distance (17.7 m.u., as com-
vg band vgb. To determine whether a particular class of pared with 12.3 and 6.4) and must therefore be the outside
progeny is parental or recombinant for vg and b, we do not genes, flanking pr in the middle (Fig. 5.10b). But as with the
care whether the flies are pr or pr. Thus, to the nearest tenth X-linked y and r genes analyzed by Sturtevant, the distance
of a map unit, the vg ↔ b distance, calculated as the per- separating the outside vg and b genes (17.7) does not equal
centage of recombinants in the total number of progeny, is the sum of the two intervening distances (12.3 6.4 18.7).
252 241 131 118 In the next section, we learn that the reason for this discrep-
100 ancy is the rare occurrence of double crossovers.
4197
17.7 m.u. 1vg 4 b distance2
Similarly, since recombinants for the vg–pr gene pair are Three-Point Crosses Allow Correction
vg pr and vg pr, the interval between these two genes is for Double Crossovers
252 241 13 9 Figure 5.11 depicts the homologous autosomes of the F1

100 females that are heterozygous for the three genes vg, pr,
4197 and b. A close examination of the chromosomes reveals the
12.3 m.u. 1vg 4 pr distance2 kinds of crossovers that must have occurred to generate the
har48464_ch05_123-166 07/19/2006 08:53 PM Page 137
CONFIRMING PAGES
5.2 Mapping: Locating Genes Along a Chromosome 137
vg pr b
(a) Parental chromosomes
vg pr b Sister chromatids
Region 1 Region 2 Homologous
chromosomes
vg+ pr + b+ of F1 females
vg+ pr + b+ Sister chromatids
(b) Crossover in region 1 Resultant chromatids

vg pr b vg pr b
vg pr b vg pr + b+
vg+ pr + b+ vg+ pr b
vg+ pr + b+ vg+ pr + b+
(c) Crossover in region 2 Resultant chromatids

vg pr b vg pr b
vg pr b vg pr b+
vg+ pr + b+ vg+ pr + b
(d) Double crossover; one crossover in each region Resultant chromatids

vg pr b vg pr b
vg pr b vg pr + b
vg+ pr + b+ vg+ pr b+
Figure 5.11 Inferring the location of a crossover event. Once you establish the order of genes involved in a three-point cross, it is
easy to determine which crossover events gave rise to particular recombinant gametes. Note that double crossovers are needed to gen-
erate gametes in which the gene in the middle has recombined relative to the parental combinations for the genes at the ends.
classes and numbers of progeny observed. In this and sub- The gametes produced by such double crossovers still have
sequent figures, the chromosomes depicted are in late the parental configuration for the outside genes vg and b, even
prophase/early metaphase of meiosis I, when there are though not one but two exchanges must have occurred.
four chromatids for each pair of homologous chromo- Because of the existence of double crossovers, the
somes. As we have suggested previously and demonstrate vg ↔ b distance of 17.7 m.u. calculated in the previous
more rigorously later, prophase I is the stage at which re- section does not reflect all of the recombination events
combination takes place. Note that we call the space be- producing the gametes that gave rise to the observed
tween vg and pr “region 1” and the space between pr and progeny. To correct for this oversight, it is necessary to
b “region 2.” adjust the recombination frequency by adding the double
Recall that the progeny from a testcross between triply crossovers twice, since each individual in the double
heterozygous F1 females and males homozygous for the re- crossover groups is the result of two exchanges between
cessive allele of all three traits fall into eight groups (re- vg and b. The corrected distance is
view Fig. 5.10). Flies in the two largest groups carry the
same configurations of genes as did their grandparents of 252 241 131 118 13 13 9 9
100
the P generation: vg b pr and vg bpr; they thus repre- 4197
sent the parental classes (Fig. 5.11a). The next two 18.7 m.u.
groups—vgb pr and vg bpr—are composed of recom-
binants that must be the reciprocal products of a crossover This value makes sense because you have accounted for all
in region 1 between vg and pr (Fig. 5.11b). Similarly the of the crossovers that occur in region 1 as well as all of the
two groups containing vg b pr and vg b pr flies must crossovers in region 2. As a result, the corrected value of
have resulted from recombination in region 2 between pr 18.7 m.u. for the distance between vg and b is now exactly
and b (Fig. 5.11c). the same as the sum of the distances between vg and pr
But what about the two smallest groups made up of rare (region 1) and between pr and b (region 2).
vg b pr and vg b pr recombinants? What kinds of chro- As previously discussed, when Sturtevant originally
mosome exchange could account for them? Most likely, they mapped several X-linked genes in Drosophila by two-point
result from two different crossover events occurring simulta- crosses, the locus of the rudimentary wings (r) gene was
neously, one in region 1, the other in region 2 (Fig. 5.11d). ambiguous. A two-point cross involving y and r gave a
har48464_ch05_123-166 07/19/2006 08:53 PM Page 138
CONFIRMING PAGES
recombination frequency of 42.9, but the sum of all the in- The fact that the number of observed double crossovers
tervening distances was 55.0 (review Fig. 5.9 on p. 135). is less than the number expected if the two exchanges are
This discrepancy occurred because the two-point cross (see independent events suggests that the occurrence of one
Fig. 5.9c) ignored double crossovers that might have crossover reduces the likelihood that another crossover
occurred in the large interval between the y and r genes. will occur in an adjacent part of the chromosome. This
The data summing the smaller intervening distances (see phenomenon—of crossovers not occurring independently—
Fig. 5.9d) accounted for at least some of these double is called chromosomal interference. Interference may
crossovers by catching recombinations of gene pairs be- exist to ensure that every pair of homologous chromo-
tween y and r. Moreover, each smaller distance is less likely somes undergoes at least one crossover event. The rea-
to encompass a double crossover than a larger distance, so soning behind this hypothesis is straightforward: It is
each number for a smaller distance is inherently more critical that every pair of homologous chromosomes sus-
accurate. Note that even a three-point cross like the one for tain one or more crossover events because such events
vg, pr, and b ignores the possibility of two recombination help the chromosomes orient properly at the metaphase
events taking place in, say, region 1. For greatest accuracy, plate during the first meiotic division. Indeed, homolo-
it is always best to construct a map using many genes gous chromosome pairs without crossovers often segre-
separated by relatively short distances. gate improperly. If only a limited number of crossovers
can occur during each meiosis and interference lowers
the number of crossovers on large chromosomes, then
Interference: The Number of Double Crossovers interference also raises the probability of crossovers
May Be Less Than Expected occurring on small chromosomes. This increases the
In a three-point cross following three linked genes, of the likelihood that at least one crossover will take place on
eight possible genotypic classes, the two parental classes every homologous pair. Though the molecular mecha-
contain the largest number of progeny, while the two dou- nism underlying interference is not yet clear, recent
ble recombinant classes, resulting from double crossovers, experiments suggest that interference is mediated by the
are always the smallest (see Fig. 5.10). We can understand synaptonemal complex.
why double-crossover progeny are the rarest by looking at Interference is not uniform and may vary even for dif-
the probability of their occurrence. If an exchange in region ferent regions of the same chromosome. Investigators can
1 of a chromosome does not affect the probability of an obtain a quantitative measure of the amount of interference
exchange in region 2, the probability that both will occur in different chromosomal intervals by first calculating a
simultaneously is the product of their separate probabilities coefficient of coincidence, defined as the ratio between the
(recall the product rule in Chapter 2, p. 21). For example, if actual frequency of double crossovers observed in an ex-
progeny resulting from recombination in region 1 alone periment and the number of double crossovers expected on
account for 10% of the total progeny (that is, if region 1 is the basis of independent probabilities.
10 m.u.) and progeny resulting from recombination in frequency observed
region 2 alone account for 20%, the probability of a double Coefficient of coincidence
frequency expected
crossover (one event in region 1, the second in region 2) is
0.10 0.20 0.02, or 2%. This makes sense because the For the three-point cross involving vg, pr, and b, the coeffi-
likelihood of two rare events occurring simultaneously is cient of coincidence is
even less than that of either rare event occurring alone.
0.52
As we have seen, in a three-point cross of the vg pr b 0.66
trio of genes, the two classes of progeny arising from 0.79
double crossovers contain the fewest progeny. The numeri- The definition of interference itself is
cal frequencies of observed double crossovers, however, do
not coincide with expectations derived from the law of the Interference 1 coefficient of coincidence
product. Let’s look at the actual numbers. The probability In this case, it is
of a single crossover between vg and pr is 0.123 (corre-
sponding to 12.3 m.u.), and the probability of a single 1 0.66 0.34
crossover between pr and b is 0.064 (6.4 m.u.). The prod-
uct of these probabilities is To understand the meaning of interference, it is helpful
to contrast what happens when there is no interference with
0.123 0.064 0.0079 0.79% what happens when it is complete. If interference is 0, the
But the observed proportion of double crossovers (see frequency of observed double crossovers equals expecta-
Fig. 5.10) was tions, and crossovers in adjacent regions of a chromosome
occur independently of each other. If interference is
13 9 complete, that is, if interference 1 because the coeffi-
100 0.52%
4197 cient of coincidence 0, no double crossovers occur in the
har48464_ch05_123-166 07/19/2006 08:53 PM Page 139
CONFIRMING PAGES
experimental progeny because one exchange effectively

w+ w y+ y m+ m X/Y
prevents another. As an example, in a particular three-point
cross in mice, the recombination frequency for the pair of Before data analysis, you do not
genes on the left (region 1) is 20, and for the pair of genes know the gene order or allele
combination on each chromosome.
on the right (region 2), it is also 20. Without interference,
the expected rate of double crossovers in this chromosomal Male progeny 2278 w+ y+ m / Y Parental class
interval is 2157 w y m+ / Y (noncrossover)
1203 w y m /Y Crossover in region 2
0.20 0.20 0.04, or 4% 1092 w+ y+ m+ / Y (between w and m)
49 w+ y m / Y Crossover in region 1
but when investigators observed 1000 progeny of this (between y and m)
41 w y+ m+ / Y
cross, they found 0 double recombinants instead of the 2 w+ y m+ / Y Double
expected 40. 1 w y+ m / Y crossovers
6823
After data analysis, you can conclude that
The Arrangement of Alleles in Double the gene order and allele combinations on the
Recombinants Indicates the Relative X chromosomes of the F1 females were y w m+ / y+ w + m.
Order of Three Genes
As we pointed out earlier, in three-point crosses of linked Figure 5.12 How three-point crosses verify Sturtevant’s
genes, the smallest of the eight possible classes of prog- map. The parental classes correspond to the two X chromosomes
eny are the two that contain double recombinants gener- in the F1 female. The makeup of the double recombinant classes
ated by double crossovers. It is possible to use the shows that w must be the gene in the middle.
composition of alleles in these double crossover classes
to determine which of the three genes lies in the middle,
even without calculating any recombination frequencies.
Consider again the progeny of a three-point testcross
looking at the vg, pr, and b genes. The F1 females are look at a three-point cross following y, w, and m, these
vg pr b / vg pr b. As Fig. 5.11d demonstrated, testcross questions disappear.
progeny resulting from double crossovers in the trihybrid Figure 5.12 tabulates the classes and numbers of
females of the F1 generation received gametes from their male progeny arising from females heterozygous for the
mothers carrying the allelic combinations vg pr b and y, w, and m genes. Since these male progeny receive their
vg pr b. In these individuals, the alleles of the vg and only X chromosome from their mothers, their pheno-
b genes retain their parental associations (vg b and vg b), types directly indicate the gametes produced by the het-
while the pr gene has recombined with respect to both erozygous females. In each row of the figure’s table, the
the other genes ( pr b and pr b; vg pr and vg pr). genes appear in an arbitrary order that does not presup-
The same is true in all three-point crosses: In those pose knowledge of the actual map. As you can see, the
gametes formed by double crossovers, the gene whose two classes of progeny listed at the top of the table out-
alleles have recombined relative to the parental configu- number the remaining six classes, which indicates that
rations of the other two genes must be the one in the all three genes are linked to each other. Moreover, these
middle. largest groups, which are the parental classes, show that
the two X chromosomes of the heterozygous females
were w y m and w y m.
Among the male progeny in Fig. 5.12, the two smallest
Comprehensive Example: Using a Three-Point
classes, representing the double crossovers, have X chro-
Cross to Reanalyze Sturtevant’s Map of the
mosomes carrying w y m and w y m combinations, in
Drosophila X Chromosome
which the w alleles are recombined relative to those of y
The technique of looking at double recombinants to dis- and m. The w gene must therefore lie between y and m, ver-
cover which gene has recombined with respect to both other ifying Sturtevant’s original assessment.
genes allows immediate clarification of gene order even in To complete a map based on the w y m three-point cross,
otherwise difficult cases. Consider the three X-linked genes you can calculate the interval between y and w (region 1)
y, w, and m that Sturtevant located in his original mapping
49 41 1 2
experiment (see Fig. 5.9 on p. 135). Since the distance be- 100 1.3 m.u.
tween y and m (34.3 m.u.) appeared slightly larger than the 6823
distance separating w and m (32.8 m.u.), he concluded that as well as the interval between w and m (region 2)
w was the gene in the middle. But because of the small dif-
ference between the two numbers, his conclusion was sub- 1203 1092 2 1
100 33.7 m.u.
ject to questions of statistical significance. If, however, we 6823
har48464_ch05_123-166 07/19/2006 08:53 PM Page 140
CONFIRMING PAGES
The genetic distance separating y and m is the sum of not significant because the 0.01 0.01 0.0001 probability
of their occurring is so small. But for genes separated by 20,
1.3 33.7 35.0 m.u.
30, or 40 m.u., the probability of double crossovers skewing
Note that you could also calculate the distance be- the data takes on greater significance. A second confounding
tween y and m directly by including double crossovers factor is the 50% limit on the recombination frequency ob-
twice, to account for the total number of recombination servable in a cross. This limit reduces the precision of RF as
events detected between these two genes. a measure of chromosomal distances. No matter how far
RF 11203 1092 49 41 2 2
apart two genes are on a long chromosome, they will never
recombine more than 50% of the time. Yet a third problem is
1 12 / 6823 100 35.0 m.u.
that recombination is not uniform even over the length of a
This method yields the same value as the sum of the two in- single chromosome: Certain “hotspots” are favored sites of
tervening distances (region 1 region 2). recombination, while other areas—often in the vicinity of
Further calculations show that interference is consider- centromeres—are “recombination deserts” in which few
able in this portion of the Drosophila X chromosome, at crossovers ever take place.
least as inferred from the set of data tabulated in Fig. 5.12. Ever since Morgan, Sturtevant, and others began
The percentage of observed double recombinants was mapping, geneticists have generated mathematical equa-
tions called mapping functions to compensate for the
3 / 6823 0.00044, or 0.044%
inaccuracies inherent in relating recombination frequen-
(rounding to the nearest thousandth of a percent), while cies to physical distances. These equations generally
the percentage of double recombinants expected on make large corrections for RF values of widely separated
the basis of independent probabilities by the law of the genes, while barely changing the map distances separat-
product is ing genes that lie close together. This reflects the fact
that multiple recombination events and the 50% limit on
0.013 0.337 0.0044, or 0.44%
recombination do not confound the calculation of dis-
Thus, the coefficient of coincidence is tances between closely linked genes. However, the cor-
rections for large distances are at best imprecise,
0.044 / 0.44 0.1
because mapping functions are based on simplifying as-
and the interference is sumptions (such as no interference) that rarely apply to
the specific situation under consideration. Thus, the best
1 0.1 0.9
way to create an accurate map is still by summing many
smaller intervals, locating widely separated genes
through linkage to common intermediaries. Because the
How Close Is the Correlation Between most accurate genetic maps are drawn from a large series
of genetic crosses, they are subject to continual refine-
a Genetic Map and Physical Reality? ment as more and more newly discovered genes are
In 1911, when Sturtevant first devised his system of included.
genetic mapping based on recombination frequencies, he Rates of recombination may differ from species to
acknowledged there was at the time no way to correlate a species. We know this because recent elucidation of the
specific gene position with a physical part of the chro- complete DNA sequences of several organisms’ genomes
mosome. Subsequent experiments, however, have shown has allowed investigators to compare the actual physical
that genetic maps do in fact accurately depict the order of distances between genes (in base pairs of DNA) with ge-
genes along chromosomes. Thus, as we will see again netic map distances. They found that in humans, a map unit
and again in the many types of experiments presented corresponds on average to about 1 million base pairs. In
throughout the remainder of this book, the order of genes yeast, however, where the rate of recombination per length
revealed by mapping techniques corresponds to the order of DNA is much higher than in humans, one map unit is ap-
of those same genes along the DNA molecule of a proximately 2500 base pairs. Thus, although map units are
chromosome. useful for estimating distances between the genes of an or-
In contrast, the actual physical distances between ganism, 1% RF can reflect very different expanses of DNA
genes—that is, the amount of DNA separating them—does in different organisms.
not always show a direct correspondence to genetic map dis- Recombination rates sometimes even vary between the
tances. This is largely because the relationship between re- two sexes of the same species. Drosophila provides an ex-
combination frequency and physical distance along a treme example: No recombination occurs during meiosis in
chromosome is not simple. One complicating factor is the ex- males. If you review the examples already discussed in
istence of double, triple, and even more crossovers. When this chapter, you will discover that they all measure
genes are separated by 1 m.u. or less, double crossovers are recombination among the progeny of doubly heterozygous
har48464_ch05_123-166 07/19/2006 08:53 PM Page 141
CONFIRMING PAGES
Drosophila females. Problem 17 on p. 161 shows how ge- that it is possible to show that any gene on a chromosome is
neticists can exploit the absence of recombination in linked to another gene on the same chromosome, the num-
Drosophila males to establish rapidly that genes far apart on ber of linkage groups equals the number of pairs of homol-
the same chromosome are indeed syntenic. ogous chromosomes in the species. Humans have 23
linkage groups, mice have 20, and fruit flies have 4
(Fig. 5.13). The total genetic distance along a chromosome,
Multiple-Factor Crosses Help Establish which is obtained by adding many short distances between
genes, may be much more than 50 m.u. For example, the
Linkage Groups two long Drosophila autosomes are both slightly more than
Genes chained together by linkage relationships are known 100 m.u. in length (Fig. 5.13), while the longest human
collectively as a linkage group. When enough genes have chromosome is approximately 270 m.u. Recall, however,
been assigned to a particular chromosome, the terms chro- that even with the longest chromosomes, pairwise crosses
mosome and linkage group become synonymous. If you can between genes located at the two ends will not produce
demonstrate that gene A is linked to gene B, B to C, C to D, more than 50% recombinant progeny.
and D to E, you can conclude that all of these genes are syn- Linkage mapping has practical applications of great
tenic. When the genetic map of a genome becomes so dense importance. For example, the Fast Forward box “Gene
X chromosome
bent wing
net veins cubitus veins
yellow body 0.0 artistaless antenna 0.0 roughoid eyes
0.0 shaven hairs
0.0 scute bristles veinlet veins
1.3 star eyes 0.2 grooveless scutellum
1.5 white eyes eyeless
4.0 held-out wings
3.0 facet eyes
5.5 echinus eyes
7.5 ruby eyes
13.7 13.0 dumpy wings
crossveinless wings
16.5 thick veins
20.0 cut wings 19.2 javelin bristles
21.0 singed bristles
26.0 sepia eyes
27.7 lozenge eyes 26.5 hairy body
33.0 vermilion eyes

36.1 miniature wings
41.0 dichaete bristles
43.0 sable body 43.2 thread arista
44.0 garnet eyes 48.5 black body 44.0 scarlet eyes

51.0 reduced bristles 48.0 pink eyes
55.0 rudimentary 54.5 purple eyes 52.0 rosy eyes
56.7 forked bristles 54.8 short bristles 58.2 stubble bristles
57.0 bar eyes 55.0 light eyes spineless bristles
58.5
59.5 fused veins 57.5 cinnabar eyes bithorax body
58.7
62.5 carnation eyes 62.0 stripe body
66.7 scabrous eyes glass eyes
66.0 bobbed hairs 63.0
67.0 vestigial wings delta veins
66.2
lobe eyes 69.5 hairless bristles
72.0
70.7 ebony body
75.5 curved wings
74.7 cardinal eyes
91.1 rough eyes
100.5 plexus wings 100.7 claret eyes

Wild type
104.5 brown eyes
106.2 minute bristles
107.0 blistered wings
Figure 5.13 Drosophila melanogaster has four linkage groups. A genetic map of the fruit fly, showing the position of many
genes influencing visible phenotypes of body morphology, including those used as examples in this chapter (highlighted in bold).
Because so many Drosophila genes have been mapped, each of the four chromosomes can be represented as a single linkage group.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 142
CONFIRMING PAGES
Gene Mapping May Lead to a Cure for Cystic Fibrosis

For 40 years after the symptoms of cystic fibrosis were first • After a review of many family pedigrees containing first-
described in 1938, there was no molecular clue—no visible cousin marriages had confirmed that cystic fibrosis is
chromosomal abnormality transmitted with the disease, no most likely determined by a single gene (the CF gene),
identifiable protein defect carried by affected individuals— investigators collected white blood cells from 47 families
suggesting the genetic cause of the disorder. As a result, with two or more affected children, obtaining genetic
there was no effective treatment for the 1 in 2000 Caucasian data from 106 patients, 94 parents, and 44 unaffected
Americans born with the disease, most of whom died before siblings.
they were 30. In the 1980s, however, geneticists were able to
combine recently invented techniques for looking directly at • They next tried to discover if any other trait is reliably
DNA with maps constructed by linkage analysis to pinpoint a transmitted with the condition. Analyses of the easily
precise chromosomal position, or locus, for the cystic fibrosis obtainable serum enzyme paroxonase showed that its
gene. Knowledge of this locus made it possible to identify gene (known as the PON gene) is indeed linked to the CF
and clone the gene, discover the protein it encodes, under- gene. At first, this knowledge was not that helpful, be-
stand the biochemical mechanisms disrupted by faulty cause the PON gene had not yet been assigned to a chro-
alleles, and develop better therapies for this life-threatening mosome.
condition. Researchers hope to use the purified cystic fibro-
sis gene as the basis for a future cure of the disease by gene • Then, in the early 1980s, geneticists developed a large se-
therapy. ries of DNA markers, based on new techniques that en-
The mappers of the cystic fibrosis gene faced an over- abled them to recognize variations in the genetic
whelming task. They were searching for a gene that encoded material. A DNA marker is a piece of DNA of known size,
an unknown protein, a gene that had not yet even been as- representing a specific locus, that comes in identifiable
signed to a chromosome. It could lie anywhere amid the variations. These allelic variations segregate according to
23 pairs of chromosomes in a human cell. Imagine looking for Mendel’s laws, which means it is possible to follow their
a close friend you lost track of years ago, who might now be transmission as you would any gene’s. For example, if
anywhere in the world. You would first have to find ways to one allelic form of DNA marker A cosegregates with the
narrow the search to a particular continent (the equivalent abnormal CF allele, while another form of marker A as-
of a specific chromosome in the gene mappers’ search); then sorts with the normal copy of the gene, these identifiable
to a country (the long or short arm of the chromosome); next variations in DNA serve as molecular landmarks located
to the state or province, county, city, or town, and street (all near the CF gene. Chapters 11 and 12 explain the discov-
increasingly narrow bands of the chromosome); and finally, ery and use of DNA markers in greater detail; for now, it
to a house address (the locus itself). Here, we briefly summa- is only important to know that they exist and can be
rize how researchers applied some of these steps in mapping identified.
the cystic fibrosis gene.
Mapping May Lead to a Cure for Cystic Fibrosis” de- ences derived from examining the phenotypes of diploid
scribes how researchers used linkage information to locate progeny resulting from random unions of random prod-
the gene for this important human hereditary disease. ucts of meiosis. For such diploid organisms, we do not
know which, if any, of the parents’ other progeny arose
from gametes created in the same meiosis. Because of this
limitation, the analysis of random products of meiosis in
Tetrad Analysis in Fungi: A Powerful Tool diploid organisms is based on statistical samplings of
for Mapping and Understanding the large populations.
In contrast, various species of fungi provide a unique
Mechanisms of Recombination opportunity for genetic analysis because they house all
With Drosophila, mice, peas, people, and other diploid four haploid products of each meiosis in a sac called an
organisms, each individual represents only one of the four ascus (plural, asci). These haploid cells, or ascospores
potential gametes generated by each parent in a single (also known as haplospores), can germinate and survive as
meiotic event. Thus, until now, our presentation of link- viable haploid individuals that grow and perpetuate them-
age, recombination, and mapping has depended on infer- selves by mitosis. The phenotype of such haploid fungi is
har48464_ch05_123-166 07/19/2006 08:53 PM Page 143
CONFIRMING PAGES
By 1986, linkage analyses of hundreds of DNA markers Chromosome 7

had shown that one marker, known as D7S15, is linked with
both the PON gene and the CF gene. Once they had estab-
lished the fact of linkage, researchers computed recombina-
tion frequencies and found that the distance from the DNA
marker to the CF gene was 15 cM; from the DNA marker to
the PON gene, 5 cM; and from PON to CF, 10 cM. After re-
peating their analyses with an ever-extending group of
families, they concluded that the order of the three loci was
D7S15-PON-CF (Fig. A). Since the CF gene could lie 15 cM in
either of two directions from the DNA marker, the area un- D7S15
5 cM
der investigation was approximately 30 cM. And since the PON
human genome consists of roughly 3000 cM, this step of 10 cM
met
linkage analysis narrowed the search to 1% of the human 1 cM
Band 7q31 CF
genome. 1 cM
J3.11
18 cM
• Next, the DNA marker D7S15 was localized to the long βTR
arm of chromosome 7, which meant that the gene for
cystic fibrosis also resides in that same region of chromo-
some 7. Researchers had now placed the CF gene in a cer- Figure A How molecular markers helped locate the gene for
tain country on a particular genetic continent. cystic fibrosis (CF).
• Finally, investigators discovered linkage with several

other markers on the long arm of chromosome 7. These
included a DNA marker known as J3.11, a gene for a key
molecule of the immune system (the
chain of the By 1989, researchers had used this mapping information
T-cell receptor), and an “oncogene” called met (a gene to identify and clone the CF gene on the basis of its location.
which has mutant alleles that contribute to cancer). Two And by 1992, they had shown it encodes a cell membrane
of the markers turned out to be separated from the CF protein that regulates the flow of chloride ions into and out
gene by a distance of only 1 cM. With all these markers, of cells (review the Fast Forward box “Genes Encode Pro-
it became possible to place the CF gene in the middle teins” in Chapter 2). This knowledge has become the basis of
third of chromosome 7’s long arm, on band 31 (band new therapies to open up ion flow (older therapies treated
7q31, Fig. A). For families with at least one child who the results of flow blockage with only minimal success), as
has cystic fibrosis, geneticists using DNA analyses of well as gene therapies to introduce normal copies of the CF
these closely linked markers could now identify carriers gene into the cells of CF patients. Although only in the early
of an abnormal copy of the CF gene with substantial stages of development, such gene therapy holds out hope of
confidence. an eventual cure for cystic fibrosis.
a direct representation of their genotype, without compli- Haploid cells of both yeast and Neurospora normally
cations of dominance. Figure 5.14 illustrates the life cy- reproduce vegetatively (that is, asexually) by mitosis.
cles of two fungal species that preserve their meiotic However, sexual reproduction is possible because the hap-
products in a sac. One, the normally unicellular baker’s loid cells come in two mating types, and cells of opposite
yeast (Saccharomyces cerevisiae), is sold in supermarkets mating types can fuse to form a diploid zygote (Fig. 5.14).
and contributes to the texture, shape, and flavor of bread; it In baker’s yeast, these diploid cells are stable and can re-
generates four ascospores with each meiosis. The other, produce through successive mitotic cycles. Stress, such as
Neurospora crassa, is a bread mold that renders the bread that caused by a scarcity or lack of essential nutrients, in-
on which it grows inedible; it too generates four as- duces the diploid cells of yeast to enter meiosis. In bread
cospores with each meiosis, but at the completion of meio- mold, the diploid zygote instead immediately undergoes
sis, each of the four haploid ascospores immediately meiosis, so the diploid state is only transient.
divides once by mitosis to yield four pairs, for a total of Mutations in haploid yeast and mold affect many dif-
eight haploid cells. The two cells in each pair of Neu- ferent traits, including the appearance of the cells and their
rospora ascospores have the same genotype, because they ability to grow under particular conditions. For instance,
arose from mitosis. yeast cells with the his4 mutation are unable to grow in the
har48464_ch05_123-166 07/19/2006 08:53 PM Page 144
CONFIRMING PAGES
(a) Saccharomyces cerevisiae Ascus containing four

haploid ascospores
α
a-mating type a α-mating type
ascospore released ascospore released
α a
a α
Germination Germination
Vegetative life Vegetative life

cycle a α cycle
(haploid) (haploid)
a
Meiosis α
Budding Budding
a/α
a/α
a/α Zygote formed
Vegetative life
cycle
(diploid)
Budding
Ascus containing 8
(b) Neurospora crassa
haploid ascospores
A-mating type ascospore a-mating type ascospore

released; germination released; germination
A a
Mitosis
Tetrad Asexual spores

Conidia Meiosis II (conidia)
Meiosis I
Germination
Germination
Vegetative life Ascus formation Vegetative life

cycle (haploid) 2n zygotes = A/a cycle (haploid)
Sexual life cycle: Sexual life cycle:

A-type cells fuse with Nuclear fusion a-type cells fuse with a
A opposite mating type opposite mating type
a A
Binucleate cell
(n + n) = A + a
Figure 5.14 The life cycles of the yeast Saccharomyces cerevisiae and the bread mold Neurospora crassa. Both S. cerevisiae
and N. crassa have two mating types that can fuse to form diploid cells that undergo meiosis. (a) Yeast cells can grow vegetatively either
as haploids or diploids. The products of meiosis in a diploid cell are four haploid ascospores that are arranged randomly in unordered
yeast asci. (b) The diploid state in Neurospora exists only for a short period. Meiosis in Neurospora is followed by mitosis, to give
eight haploid ascospores in the ascus. The ordered arrangement of spores in Neurospora asci reflects the geometry of the meiotic and
mitotic spindles. The photograph showing a budding (mitotically dividing) yeast cell in part (a) is at much higher magnification that the
photograph displaying Neurospora asci in part (b).
har48464_ch05_123-166 07/19/2006 08:53 PM Page 145
CONFIRMING PAGES
absence of the amino acid histidine, while yeast with the his4 / HIS4; trp1 / TRP1, as shown in Fig. 5.15a. (In ge-
trp1 mutation cannot grow without an external source of netic nomenclature, a semicolon [;] is usually employed
the amino acid tryptophan. Geneticists who specialize in to separate genes on nonhomologous chromosomes.)
the study of yeast have devised a system of representing When conditions promote meiosis, the two unlinked
genes that is slightly different from the ones for Drosophila genes will assort independently to produce equal fre-
and mice. They use capital letters (HIS4) to designate dom- quencies of two different kinds of tetrads. In one kind, all
inant alleles and lowercase letters (his4) to represent reces- the spores are parental in that the genotype of each spore
sives. For most of the yeast genes we will discuss, the is the same as one of the parents: his4 TRP1 or HIS4 trp1
wild-type alleles are dominant and may be represented by (Fig. 5.15b). A tetrad that contains four parental class
the alternative shorthand “”, while the symbol for the re- haploid cells is known as a parental ditype (PD). Note
cessive alleles remains the lowercase abbreviation (his4). that di-, meaning two, indicates there are two possible
Remember, however, that dominance or recessiveness is parental combinations of alleles; the PD tetrad contains
relevant only for diploid yeast cells, not for haploid cells two of each combination. The second kind of tetrad, aris-
that carry only one allele. ing from the equally likely alternative distribution of
After meiosis, the assemblage of four ascospores (or chromosomes during meiosis, contains four recombinant
four pairs of ascospores) in a single ascus is called a spores: two his4 trp1 and two HIS4 TRP1 (Fig. 5.15c).
tetrad. (Note that this is a second meaning for the term This kind of tetrad is termed a nonparental ditype
tetrad. In Chapter 4, a tetrad was the four homologous (NPD), because the two parental classes have recom-
chromatids—two in each chromosome of a bivalent— bined to form two reciprocal nonparental combinations
synapsed during the prophase and metaphase of meiosis of alleles.
I. Here, it is the four products of a single meiosis held A third kind of tetrad also appears when his4/HIS4;
together in a sac. Since the four chromatids of a biva- trp1/TRP1 cells undergo meiosis. Called a tetratype (T)
lent give rise to the four products of meiosis, the two from the Greek word for “four,” it carries four kinds of hap-
meanings of tetrad refer to almost the same things.) loid cells: two different parental class spores (one his4
In yeast, each tetrad is unordered; that is, the four TRP1 and one HIS4 trp1) and two different recombinants
meiotic products, known as spores, are arranged at ran- (one his4 trp1 and one HIS4 TRP1). Tetratypes result from
dom within the ascus. In Neurospora crassa, each tetrad a crossover between one of the two genes and the
is ordered, with the four pairs, or eight haplospores, centromere of the chromosome on which it is located
arranged in a line. (Fig. 5.15d).
To analyze both unordered and ordered tetrads, re- Figure 5.15e displays the data from one experiment.
searchers can release the spores of each ascus, induce the Bear in mind that the column headings of PD, NPD, and T
haploid cells to germinate under appropriate conditions, refer to tetrads (the group of four cells produced in meio-
and then analyze the genetic makeup of the resulting hap- sis) and not to individual haploid cells. Because the spores
loid cultures. The data they collect in this way enable them released from a yeast ascus are not arranged in any partic-
to identify the four products of a single meiosis and com- ular order, the order in which the spores are listed does not
pare them with the four products of many other distinct matter. The classification of a tetrad as PD, NPD, or T is
meioses. Ordered tetrads offer another possibility. With the based solely on the number of parental and recombinant
aid of a dissecting microscope, investigators can recover spores found in the ascus.
the ascospores in the order in which they occur within the
ascus and thereby obtain additional information that is use-
When PD NPD, Two Genes Are Unlinked
ful for mapping. We look first at the analysis of randomly
arranged spores, using the unordered tetrads of yeast as an A cross following two unlinked genes must give equal
example. We then describe the additional information that numbers of individual parental and recombinant spores.
can be gleaned from the microanalysis of ordered tetrads, This is simply another way of stating Mendel’s second
using Neurospora as our model organism. law of independent assortment, which predicts a 50% re-
combination frequency in such cases. Since T tetrads, re-
gardless of their number, contain two recombinant and
Tetrads Can Be Characterized by the Number of
two nonrecombinant spores and since all four spores in
Parental and Recombinant Spores They Contain
PD tetrads are parental, the only way 50% of the total
What kinds of tetrads arise when diploid yeast cells het- progeny spores could be recombinant (as demanded by
erozygous for two genes on different chromosomes are independent assortment) is if the number of NPDs (with
induced to undergo meiosis? Consider a mating between four recombinant spores apiece) is the same as the
a haploid strain of yeast of mating type a, carrying the number of PDs. For this reason, if PD NPD (as in
his4 mutation and the wild-type allele of the TRP1 gene, Fig. 5.15e), the two genes must be unlinked, either be-
and a strain of the opposite mating type that has the cause they reside on different chromosomes or because
genotype HIS4 trp1. The resulting a / diploid cells are they lie very far apart on the same chromosome.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 146
CONFIRMING PAGES
(a) his4 TRP1 HIS4 trp1

Parents
a α
his4 TRP1
Diploid cell HIS4 trp1
a/α
Meiosis I Meiosis II Tetrad Ascus
(b)
his4 TRP1
his4 TRP1
his4 TRP1 his4 TRP1 his4 TRP1 his4 HIS4

TRP1 trp1
his4 TRP1
HIS4 trp1 his4 HIS4

HIS4 trp1
HIS4 trp1 TRP1 trp1
HIS4 trp1
HIS4 trp1 HIS4 trp1
Parental ditype (PD)
(c) his4 trp1

his4 trp1
his4 trp1 his4 trp1 his4 trp1 his4 HIS4

trp1 TRP1
his4 trp1
HIS4 TRP1 HIS4 his4

HIS4 TRP1
HIS4 TRP1 TRP1 trp1
HIS4 TRP1
HIS4 TRP1 HIS4 TRP1
Nonparental ditype (NPD)
(d) his4 TRP1

his4 TRP1
his4 TRP1 HIS4 TRP1 HIS4 TRP1 HIS4 his4
TRP1 TRP1
his4 TRP1
HIS4 trp1
his4 trp1 HIS4 his4
HIS4 trp1 his4 trp1 trp1 trp1
HIS4 trp1 HIS4 trp1

Tetratype (T)
(e) PD NPD T
HIS4 trp1 his4 trp1 his4 trp1
HIS4 trp1 his4 trp1 his4 TRP1
his4 TRP1 HIS4 TRP1 HIS4 trp1
his4 TRP1 HIS4 TRP1 HIS4 TRP1
Number of tetrads 31 28 41
Figure 5.15 How meiosis can generate three kinds of tetrads when two genes are on different chromosomes. (a) Parental
cross. (b) and (c) In the absence of recombination, the two equally likely alternative arrangements of two pairs of chromosomes
yield either PD or NPD tetrads. T tetrads are made only if either gene recombines with respect to its corresponding centromere, as in
(d). Numerical data in (e) show that the number of PD tetrads the number of NPD tetrads when the two genes are unlinked.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 147
CONFIRMING PAGES
P a r g3 u r a 2 ARG 3 U R A 2 Duplication Meiosis I Meiosis II

(a-mating type) (α-mating type)
(a) No crossing-over (NCO)
Diploid cell a r g3 u r a 2 / ARG 3 U R A 2
arg 3 ura 2
arg 3 ura 2 arg 3 ura 2
Meiosis
arg 3 ura 2
ARG 3 URA 2 ARG 3 URA 2 ARG 3 URA 2

Products of PD NPD T
ARG 3 URA 2
meiosis a r g3 u r a 2 a r g3 URA2 a r g3 ura2 ARG 3 URA 2 ARG 3 URA 2
a r g3 u r a 2 a r g3 URA2 a r g3 URA2 Parental ditype
ARG 3 U R A 2 ARG 3 ur a 2 ARG 3 ura2
ARG 3 U R A 2 ARG 3 ur a 2 ARG 3 URA2 (b) Single crossover (SCO)
Number of 127 3 70
tetrads arg 3 ura 2
arg 3 URA 2
arg 3 ura 2 arg 3 URA 2
Figure 5.16 When genes are linked, PDs exceed NPDs.
ARG 3 ura 2
ARG 3 URA 2 ARG 3 ura 2
ARG 3 URA 2
ARG 3 URA 2 ARG 3 URA 2
Tetratype
When PDs Significantly Outnumber NPDs, (c) Double crossover (DCO)
Two Genes Are Linked 2-strand
The genetic definition of linkage is the emergence of more arg 3 ura 2 arg 3 ura 2
arg 3 ura 2
parental types than recombinants among the progeny of a arg 3 ura2 arg 3 ura 2 arg 3 ura 2
doubly heterozygous parent. In the preceding section, we ARG 3 URA 2
saw that tetratypes always contribute an equal number of ARG 3 URA 2
parental and recombinant spores. Thus, with tetrads, link- ARG 3 URA 2 ARG 3 URA 2
age exists only when PD NPD, that is, when the num- Parental ditype
ber of PD tetrads (carrying only parental type spores) (d) DCO
substantially exceeds the number of NPD tetrads (contain- 3-strand
ing only recombinants). By analyzing an actual cross in- arg 3 ura 2 arg 3 ura 2
arg 3 ura 2
volving linked genes, we can see how this follows from the arg 3 ura 2 arg 3 URA 2
arg 3 URA 2
events occurring during meiosis. ARG 3 URA 2
A haploid yeast strain containing the arg3 and ura2
ARG 3 ura 2
mutations was mated to a wild-type ARG3 URA2 haploid ARG 3 URA 2 ARG 3 ura 2
strain (Fig. 5.16). When the resultant a / diploid was in- Tetratype
duced to sporulate (that is, undergo meiosis), the 200 tetrads (e) DCO
produced had the distribution shown in Fig. 5.16. As you 3-strand
can see, the 127 PD tetrads far outnumber the 3 NPD arg 3 ura 2 arg 3 URA 2
arg 3 URA 2
tetrads, suggesting that the two genes are linked. arg 3 ura 2
Figure 5.17 shows how we can explain the particular ARG 3 ura 2
ARG 3 URA 2 ARG 3 ura 2
kinds of tetrads observed in terms of the various types of
ARG 3 URA 2
crossovers that could occur between the linked genes. If no ARG 3 URA 2 ARG 3 URA 2
crossing-over occurs between the two genes, the resulting Tetratype

tetrad must be PD; since none of the four chromatids par- (f) DCO
ticipates in an exchange, all of the products are of parental 4-strand
configuration (Fig. 5.17a). A single crossover between arg 3 URA 2
arg 3 ura 2 arg 3 URA 2
ARG3 and URA2 will generate a tetratype, containing four arg 3 URA 2
genetically different spores (Fig. 5.17b). But what about arg 3 ura 2 arg 3 URA 2
ARG 3 URA 2 ARG 3 ura 2 ARG 3 ura 2
double crossovers? There are actually four different possi-
ARG 3 ura 2
bilities, depending on which chromatids participate, and ARG 3 URA 2 ARG 3 ura 2
each of the four should occur with equal frequency. A dou- Nonparental ditype
ble crossover involving only two chromatids (that is, one Figure 5.17 How crossovers between linked genes generate
where both crossovers affect the same two chromatids) different tetrads. (a) PDs arise when there is no crossing-over.
produces only parental-type progeny, generating a PD (b) Single crossovers between the two genes yield tetratypes.
tetrad (Fig. 5.17c). Three-strand double crossovers can (c) to (f) Double crossovers between linked genes can generate
occur in the two ways depicted in Fig. 5.17d and e; either PD, T, or NPD tetrads, depending on which chromatids participate
way, a tetratype results. Finally, if all four strands take in the crossovers.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 148
CONFIRMING PAGES
part in the two crossovers (one crossover involves two

Recombination Duplication Meiosis I Meiosis II
strands and the other crossover, the other two strands), all
four progeny spores will be recombinant, and the result- a b+
ing tetrad is NPD (Fig. 5.17f). Thus, if two genes are a b+
a b+
linked, the only way to generate an NPD tetrad is through a b a b+ a b+
a four-strand double exchange. Meioses with crossovers a b+
generating such a specific kind of double recombination a+ b a+ b

a+ b+ a+ b
must be a lot rarer than no crossing-over or single crossovers, a+ b
which produce PD and T tetrads, respectively. This ex- a+ b a+ b
plains why, if two genes are linked, PD must greatly Nonparental

exceed NPD. ditype
Figure 5.18 A mistaken model: Recombination before

chromosome replication. If recombination occurred before the
Calculating the Recombination Frequency
chromosomes duplicated and if two genes were linked, most
Because we know that all of the spores in an NPD tetrad are tetrads containing recombinant spores would be NPDs instead of
recombinant and half of the four spores in a tetratype are Ts. Actual results show that the opposite is true.
recombinant, we can say that
NPD 12T
RF 100
Total tetrads
took place before chromosome duplication, every single
For the ARG3 URA2 example in Fig. 5.16,
crossover event would yield four recombinant chromatids
3 11221702 and generate an NPD tetrad (Fig. 5.18). A model assuming
RF 100 19 m.u. that recombination occurs when there are two rather than
200
four chromatids per pair of homologous chromosomes
It is reassuring that this formula gives the exact same re- would thus predict more NPD than T tetrads, but experi-
sult as calculating the RF as the percentage of individual mental observations show just the opposite (see Figs. 5.15e
recombinant spores. For example, the 200 tetrads ana- and 5.16).
lyzed in this experiment contain 800 (that is, 200 4) in- The fact that recombination takes place after the chro-
dividual spores; each NPD ascus holds 4 recombinant mosomes have replicated explains the 50% limit on re-
ascospores, and each T tetrad contains 2 recombinants. combination for genes on the same chromosome. Single
Thus, crossovers between two genes generate T tetrads contain-
14 32 12 702 ing two out of four spores that are recombinant. Thus,
RF 100 19 m.u. even if one crossover occurred between two such genes in
800
every meiosis, the observed recombination frequency
The formula used here for calculating the RF is very would be 50%. The four kinds of double crossovers de-
accurate for genes separated by small distances, but it is picted in Fig. 5.17 yield either PD tetrads with 0/4 recom-
less reliable for more distant genes because it does not binants; T tetrads carrying 2/4 recombinants; differently
account for all types of double crossovers. Problem 35 on derived T tetrads, still carrying 2/4 recombinants; and
p. 165 will allow you to derive an alternative equation NPD tetrads with 4/4 recombinants. Because these four
that yeast geneticists often use to measure large distances kinds of double crossovers almost always occur with equal
more accurately. frequency, no more than 50% of the progeny resulting
from double (or, in fact, triple or more) crossovers can be
recombinant.
Tetrad Analysis Confirms That Recombination
Occurs at the Four-Strand Stage
Tetrad Analysis Also Demonstrates That
Both T and NPD tetrads contain recombinant spores, and
Recombination Is Usually Reciprocal
when tetrad analysis reveals linked genes, the T tetrads al-
ways outnumber the NPDs, as in the example we have been Suppose you are following linked genes A and B in a cross
discussing. This makes sense, because all single and some between A B and a b strains of yeast. If the recombination
double crossovers yield tetratypes, while only 1/4 of the that occurs during meiosis is reciprocal, every tetrad with re-
rare double crossovers produce NPDs. The very low num- combinant progeny should contain equal numbers of both
ber of NPDs establishes that recombination occurs after classes of recombinants. Observations have in general con-
the chromosomes have replicated, when there are four firmed this prediction: Every T tetrad carries one A b and one
chromatids for each pair of homologs. If recombination a B spore, while every NPD tetrad contains two of each type
har48464_ch05_123-166 07/19/2006 08:53 PM Page 149
CONFIRMING PAGES
Mitosis Resulting
Meiosis I Meiosis II Metaphase Octad
2:2
2:2 Spindle
3:1
Genetically
identical cells
Figure 5.19 In rare tetrads, the two alleles of a gene do Figure 5.20 How ordered tetrads form. Spindles form
not segregate 2:2. Researchers sporulated a HIS4 / his4 diploid parallel to the long axis of the growing Neurospora ascus, and
yeast strain and dissected the four haploid spores from three the cells cannot slide around each other. The order of ascospores
different tetrads. They then plated these spores on petri plates thus reflects meiotic spindle geometry. After meiosis, each
containing medium without histidine. Each row on the petri plate haploid cell undergoes mitosis, producing an eight-cell ascus (an
presents the four spores of a single tetrad. The top two rows octad). The octad consists of four pairs of cells; the two cells of
show the normal 2:2 segregation of the two alleles of a single each pair are genetically identical.
gene: two of the spores are HIS4 and form colonies, whereas the
other two spores are his4 and cannot grow into colonies. The
bottom row displays a rare tetrad with an unusual segregation of
3 HIS4 : 1 his4. markers, information that you cannot normally obtain
from unordered yeast tetrads. As described earlier, imme-
diately after specialized haploid Neurospora cells of dif-
ferent mating types fuse at fertilization, the diploid zygote
of recombinant. We can thus conclude that meiotic recombi- undergoes meiosis within the confines of a narrow ascus
nation is almost always reciprocal, generating two homolo- (review Fig. 5.14b on p. 144). At the completion of meio-
gous chromosomes that are inverted images of each other. sis, each of the four haploid meiotic products divides once
There are, however, exceptions. Very rarely, a particu- by mitosis, yielding an octad of eight haploid ascospores.
lar cross produces tetrads containing unequal numbers of Dissection of the ascus at this point allows one to deter-
reciprocal classes, and such tetrads cannot be classified as mine the phenotype of each of the eight haploid cells.
PD, NPD, or T. In these exceptional tetrads, the two input The cross-sectional diameter of the ascus is so small
alleles of one of the genes, instead of segregating at a ratio that cells cannot slip past each other. Moreover, during
of 2A : 2a, produce ratios of 1A : 3a or 3A : 1a or even 0A: each division after fertilization, the microtubule fibers of
4a or 4A : 0a (Fig. 5.19). In these same tetrads, markers the spindle extend outward from the centrosomes parallel
such as B/b and C/c that flank the A or a allele on the same to the long axis of the ascus. These facts have two impor-
chromosome still segregate 2B : 2b and 2C : 2c. Moreover, tant repercussions. First, when each of the four products of
careful phenotypic and genetic tests show that even when meiosis divides once by mitosis, the two genetically identi-
alleles do not segregate 2:2, only the original two input cal cells that result lie adjacent to each other (Fig. 5.20).
alleles occur in the progeny. Thus, recombination, no Because of this feature, starting from either end of the
matter what ratios it generates, does not create new alleles. ascus, you can count the octad of ascospores as four cell
Geneticists believe that the unusual non-2:2 segregation pairs and analyze it as a tetrad. Second, from the precise
ratios observed in rare instances result from molecular positioning of the four ascospore pairs within the ascus,
events at the site of recombination. We discuss these you can infer the arrangement of the four chromatids of
events at the molecular level in Chapter 6. For now, it is each homologous chromosome pair during the two meiotic
simply necessary to know that the unusual ratios exist but divisions.
are quite rare. To understand the genetic consequences of the geom-
etry of the ascospores, it is helpful to consider what kinds
of tetrads you would expect from the segregation of two
Ordered Tetrads Help Locate Genes
alleles of a single gene. (In the following discussion, you
in Relation to the Centromere
will see that Neurospora geneticists denote alleles with
Analyses of ordered tetrads, such as those produced by the symbols similar to those used for Drosophila, as detailed
bread mold Neurospora crassa, allow you to map the in the nomenclature guide on p. N-1 of the Appendix.) The
centromere of a chromosome relative to other genetic mutant white-spore allele (ws) alters ascospore color from
har48464_ch05_123-166 07/19/2006 08:53 PM Page 150
CONFIRMING PAGES
Meiosis I Meiosis II
Mitosis Segregation Pattern of Ascospores
(first division) (second division)
(a) First-division ws+

segregation patterns ws+
ws+ ws+
ws+ ws+
ws+ ws+
ws+ ws+
ws
ws ws
ws ws
ws ws
ws ws
or
ws
(b) Second-division
segregation patterns ws+
ws+
ws+ ws+
ws+ ws
ws ws
ws
ws+
ws ws+
ws+
ws ws+ ws+
ws ws ws
or or or
ws
Figure 5.21 Two segregation patterns in ordered asci. (a) In the absence of a crossover between a gene and its centromere,
the two alleles of a gene will separate at the first meiotic division. The result is a first-division segregation pattern in which each allele
appears in spores located on only one side of an imaginary line through the middle of the ascus. (b) A crossover between a gene and
its centromere produces a second-division segregation pattern in which both alleles appear on the same side of the middle line (but in
different spores).
wild-type black to white. In the absence of recombination, ascospores 4 and 5, because cells with only one kind of
the two alleles (ws and ws) separate from each other at allele do not arise until the end of the second meiotic
the first meiotic division because the centromeres to which division. Octads carrying this configuration of spores
they are attached separate at that stage. The second mei- display a second-division segregation pattern.
otic division and subsequent mitosis create asci in which Since second-division segregation patterns result from
the top four ascospores are of one genotype (for instance meioses in which there has been a crossover between a gene
ws) and the bottom four of the other (ws). Whether the and its centromere, the relative number of asci with this
top four are ws and the bottom four ws, or vice versa, pattern can be used to determine the gene ↔ centromere
depends on the random metaphase I orientation of the distance. In an ascus showing second-division segregation,
homologs that carry the gene relative to the long axis of one-half of the ascospores are derived from chromatids that
the developing ascus. have exchanged parts, while the remaining half arise from
The segregation of two alleles of a single gene at the chromatids that have not participated in crossovers leading
first meiotic division is thus indicated by an ascus in which to recombination. To calculate the distance between a gene
an imaginary line drawn between the fourth and the fifth and its centromere, you therefore simply divide the percent-
ascospores of the octad cleanly separates haploid products age of second-division segregation octads by 2. Geneticists
bearing the two alleles. Such an ascus displays a first- use information about the location of centromeres to make
division segregation pattern (Fig. 5.21a). more accurate genetic maps as well as to study the structure
Suppose now that during meiosis I, a crossover occurs and function of centromeres.
in a heterozygote between the white-spore gene and the
centromere of the chromosome on which it travels. As
A Numerical Example of Ordered-
Fig. 5.21b illustrates, this can lead to four equally possible
Tetrad Analysis
ascospore arrangements, each one depending on a particular
orientation of the four chromatids during the two meiotic In one experiment, a thrarg wild-type strain of Neuros-
divisions. In all four cases, both ws and ws spores are pora was crossed with a thr arg double mutant. The thr mu-
found on both sides of the imaginary line drawn between tants cannot grow in the absence of the amino acid threonine,
har48464_ch05_123-166 07/19/2006 08:53 PM Page 151
CONFIRMING PAGES
Referring again to Fig. 5.22, we find that groups A and G

(a) A Neurospora cross
are PD, because all the ascospores show parental combi-
Tetrad nations, while groups E and F, with four recombinant
A B C D E F G
group
spores, are NPD. PD is thus 72 1 73, while NPD is
Segregation thr arg thr arg thr arg thr arg+ thr arg + thr arg + thr arg 1 2 3. From these data, we can conclude that the two
pattern thr arg thr + arg thr arg + thr +arg thr +arg thr arg + thr +arg +
genes are linked.
thr +arg + thr + arg + thr + arg thr +arg + thr +arg thr +arg thr +arg +
thr +arg + thr arg+ thr + arg + thr arg thr arg + thr +arg thr arg
What is the map distance between thr and arg? For this
Total in 72 16 11 2 2 1 1
calculation, we need to find the numbers of T and NPD
group tetrads. Tetratypes are found in groups B, C, and D, and we
already know that groups E and F carry NPDs. Using the
(b) Corresponding genetic map
same formula for map distances as the one previously used
arg 16.7 m.u. thr for yeast,
7.6 m.u. 10 m.u. NPD 12T
Figure 5.22 Genetic mapping by ordered-tetrad analysis:
RF 100
Total tetrads
An example. (a) In ordered-tetrad analysis, tetrad classes are
defined not only as PD, NPD, or T but also according to whether we get
3 1122 116 11 22
they show a first- or second-division segregation pattern. Each
entry in this table represents a pair of adjacent, identical spores
in the actual Neurospora octad. (b) Genetic map derived from the
RF 100 16.7 m.u.
105
data in part (a). Ordered-tetrad analysis allows determination of
the centromere’s position as well as distances between genes. Since the distance between thr and arg is larger than
that separating either gene from the centromere, the cen-
tromere must lie between thr and arg, yielding the map in
Fig. 5.22b. The distance between the two genes calculated
while arg mutants cannot grow without a source of the by the formula above (16.7 m.u.) is smaller than the sum of
amino acid arginine; cells carrying the wild-type alleles of the two gene ↔ centromere distances (10.0 7.6 17.6
both genes can grow in medium that contains neither m.u.) because the formula does not account for all of the
amino acid. From this cross, 105 octads, considered here as double crossovers. As always, calculating map positions
tetrads, were obtained. These tetrads were classified in for more genes with shorter distances between them pro-
seven different groups—A, B, C, D, E, F, and G—as shown duces the most accurate picture.
in Fig. 5.22a. For each of the two genes, we can now find Table 5.3 summarizes some of the procedures for
the distance between the gene and the centromere of the mapping genes in fungi producing ordered and unordered
chromosome on which it is located. tetrads.
To do this for the thr gene, we count the number of
tetrads with a second-division segregation pattern for that
gene. Drawing an imaginary line through the middle of the
tetrads, we see that those in groups B, D, E, and G are the
result of second-division segregations for thr, while the TABLE 5.3 Rules for Tetrad Analysis
remainder show first-division patterns. The centromere ↔ thr
For Ordered and Unordered Tetrads
distance is thus
Considering genes two at a time, assign tetrads as PD, NPD, or T.
Percentage of second-division patterns
If PD NPD, the two genes are genetically linked.
1122 116 2 2 12
100 10 m.u. If PD NPD, the two genes are genetically independent
105 (unlinked).
Similarly, the second-division tetrads for the arg gene are The map distance between two genes if they are genetically
in groups C, D, E, and G, so the distance between arg and linked
its centromere is NDP 1122T
100
1122111 2 2 12 Total tetrads
100 7.6 m.u.
105 For Ordered Tetrads Only
The map distance between a gene and its centromere
To ascertain whether the thr and arg genes are
linked, we need to evaluate the seven tetrad groups in a (12) (# of tetrads showing second-division
different way, looking at the combinations of alleles for segregation for this gene)
100
the two genes to see if the tetrads in that group are PD, Total tetrads
NPD, or T. We can then ask whether PD NPD.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 152
CONFIRMING PAGES
5.3 Mitotic Recombination Can

Produce Genetic Mosaics
The recombination of genetic material is a critical feature of
meiosis. The chiasmata that mark recombination sites keep
homologous chromosomes together, allowing them to reach
the metaphase plate and subsequently to segregate to oppo- Single yellow spot Twin spot Single singed spot
site spindle poles during meiosis I. Moreover, recombination
of genes during meiosis helps ensure the genetic diversity of Figure 5.23 Twin spots: A form of genetic mosaicism. In a
y sn ysn Drosophila female, most of the body is wild type, but
the gametes. It is thus not surprising that eukaryotic organ-
aberrant patches showing either yellow color or singed bristles
isms express a variety of enzymes (described in Chapter 6)
sometimes occur. In some cases, yellow and singed patches are
that specifically initiate meiotic recombination. Recombina- adjacent to each other, a configuration known as twin spots.
tion can also occur during mitosis. Unlike what happens in
Transient pairing
Mitotic metaphase Daughter cells
during mitosis
(a) Crossing-over between sn and the centromere

Yellow
sn+ y sn+ y
sn+ y sn+ y
Twin
y+ sn y+ spot
sn
sn+ y sn y+ sn y+
or Singed
Wild type
sn y+ sn+ y sn y+
sn+ y sn y+
Normal
tissue
sn y+ sn+ y
sn y+ sn+ y
Wild type
(b) Crossing-over between sn and y

Yellow
sn+ y sn y Yellow
sn+ y sn y spot
sn+ y+ sn y+ Normal
sn+ y
sn+ y+ sn y+ tissue
or Wild type
Wild type
sn y+ sn+ y sn y+
sn+ y sn y+
Normal
tissue
sn+ y+ sn y
sn+ y+ sn y
Wild type

Figure 5.24 Mitotic crossing-over. (a) In a y sn y sn Drosophila female, a mitotic crossover between the centromere and sn can
produce two daughter cells, one homozygous for y and the other homozygous for sn, that can develop into adjacent aberrant patches
(twin spots). This outcome depends on a particular distribution of chromatids at anaphase (top). If the chromatids are arranged in the
equally likely opposite orientation, only phenotypically normal cells will result (bottom). (b) Crossovers between sn and y can generate
single yellow patches. In contrast, a single mitotic crossover in these females cannot produce a single singed spot if the sn gene is closer
to the centromere than the y gene. See if you can demonstrate this fact.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 153
CONFIRMING PAGES
5.3 Mitotic Recombination Can Produce Genetic Mosaics 153
meiosis, however, mitotic crossovers are initiated by mistakes Diploid yeast cells that are heterozygous for one or
in chromosome replication or by chance exposures to radia- more genes exhibit mitotic recombination in the form of
tion that break DNA molecules, rather than by a well-defined sectors: portions of a growing colony that have a different
cellular program. As a result, mitotic recombination is a rare genotype than the remainder of the colony. If a diploid yeast
event, occurring no more frequently than once in a million cell of genotype ADE2 / ade2 is placed on a petri plate, its
somatic cell divisions. Nonetheless, the growth of a colony of mitotic descendents will grow into a colony. Usually, such
yeast cells or the development of a complex multicellular colonies will appear white because the dominant wild-type
organism involves so many cell divisions that geneticists can ADE2 allele specifies that color. However, many colonies
routinely detect these rare mitotic events. will contain red sectors of diploid ade2 / ade2 cells, which
Curt Stern, an early Drosophila geneticist, originally in- arose as a result of mitotic recombination events between
ferred the existence of mitotic recombination from observa- the ADE2 gene and its centromere (Fig. 5.25). (Homozy-
tions of “twin spots” in a few fruit flies. Twin spots are gous ADE2 / ADE2 cells will also be produced by the same
adjacent islands of tissue that differ both from each other and event, but they cannot be distinguished from heterozygotes
from the tissue surrounding them. The distinctive patches because both types of cells are white.) The size of the red
arise from homozygous cells with a recessive phenotype sectors indicates when mitotic recombination took place. If
growing amid a generally heterozygous cell population dis- they are large, it happened early in the growth of the colony,
playing the dominant phenotype. In Drosophila, the yellow giving the resulting daughter cells a long time to proliferate;
(y) mutation changes body color from normal brown to yel- if they are small, it happened later.
low, while the singed bristles (sn) mutation causes body bris- Mitotic recombination is significant both as an experi-
tles to be short and curled rather than long and straight. Both mental tool and because of the phenotypic consequences of
of these genes are on the X chromosome. In 1936, Stern ex- particular mitotic crossovers. Problem 40 on p. 166 at the end
amined Drosophila females of genotype y sn / y sn. These of this illustrates how geneticists use mitotic recombination
heterozygotes were generally wild type in appearance, but to obtain information for mapping genes relative to each
Stern noticed that some flies carried patches of yellow body other and to the centromere. Mitotic crossing-over has also
color, others had small areas of singed bristles, and still oth- been of great value in the study of development because it can
ers displayed twin spots: adjacent patches of yellow cells generate animals in which different cells have different geno-
and cells with singed bristles (Fig. 5.23). He assumed that types (see Chapter 20). Finally, as the Genetics and Society
mistakes in the mitotic divisions accompanying fly develop- box “Mitotic Recombination and Cancer Formation” ex-
ment could have led to these mosaic animals containing tis- plains, mitotic recombination can have major repercussions
sues of different genotypes. Individual yellow or singed for human health.
patches could arise from chromosome loss or by mitotic
nondisjunction. These errors in mitosis would yield XO cells
containing only y (but not y) or sn (but not sn) alleles;
such cells would show one of the recessive phenotypes.
The twin spots must have a different origin. Stern rea-
soned that they represented the reciprocal products of mitotic
crossing-over between the sn gene and the centromere. The
mechanism is as follows. During mitosis in a diploid cell,
after chromosome duplication, homologous chromosomes
occasionally—very occasionally—pair up with each other.
While the chromosomes are paired, nonsister chromatids
(that is, one chromatid from each of the two homologous
chromosomes) can exchange parts by crossing-over. The
pairing is transient, and the homologous chromosomes soon
resume their independent positions on the mitotic metaphase
plate. There, the two chromosomes can line up relative to
each other in either of two ways (Fig. 5.24a). One of these
orientations would yield two daughter cells that remain het-
erozygous for both genes and thus be indistinguishable from
the surrounding wild-type cells. The other orientation, how-
ever, will generate two homozygous daughter cells, one
y sn /y sn, the other ysn /ysn. Since the two daughter
cells would lie next to each other, subsequent mitotic divi-
sions would produce adjacent patches of y and sn tissue (that
is, twin spots). Note that if crossing-over occurs between sn Figure 5.25 Mitotic recombination during the growth of
and y, single spots of yellow tissue can form, but a reciprocal diploid yeast colonies can create sectors. Arrows point to large,
singed spot cannot be generated in this fashion (Fig. 5.24b). red ade2 / ade2 sectors formed from ADE2 / ade2 heterozygotes.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 154
CONFIRMING PAGES
Mitotic Recombination and Cancer Formation

In humans, some tumors, such as those found in retinoblas- must convert an RB allele to RB, while the second could be
toma, may arise as a result of mitotic recombination. Recall a mitotic recombination producing daughter cells that be-
from our discussion of penetrance and expressivity in Chap- come cancerous because they are homozygous for the newly
ter 3 that retinoblastoma is the most malignant form of eye mutant, nonfunctional allele.
cancer. In the United States, a genetic predisposition to the Interestingly, the role of mitotic recombination in the for-
cancer is seen in about 1 in 20,000 live births. The retinoblas- mation of retinoblastoma helps explain the incomplete pene-
toma gene (RB) resides on chromosome 13, where the nor- trance and variable expressivity of the disease. People born as
mal wild-type allele (RB) encodes a protein that regulates RB/RB heterozygotes may or may not develop the condition
retinal growth and differentiation. Cells in the eye need at (incomplete penetrance). If, as usually happens, they do, they
least one copy of the normal wild-type allele to maintain may have it in one or both eyes (variable expressivity). It all de-
control over cell division. A mutation in RB leading to loss of pends on whether and in what cells of the body mitotic recom-
normal gene function is designated RB. If a cell loses both bination (or some other “homozygosing” event that affects
copies of RB, it loses regulatory control, and a tumor results. chromosome 13) occurs.
The normal, wild-type RB allele is thus known as a tumor-
suppressor gene.
People with a genetic predisposition to retinoblastoma
Transient Pairing
are born with only one functional copy of the normal RB of Homologous Mitotic Daughter
allele; their second chromosome 13 carries either a non- Chromosomes 13 Metaphase Cells
functional RB allele or no RB gene at all. If a mutagen During Mitosis
(such as radiation) or a mistake in gene replication or seg-
Normal
regation destroys or removes the single remaining normal
RB+
copy of the gene in a retinal cell in either eye, a retinoblas- RB+
toma tumor will develop at that site. In one study of people RB+
RB+ RB+
with a genetic predisposition to retinoblastoma, cells taken RB+
from eye tumors were RB homozygotes, while white RB-
RB- RB- RB-
blood cells from the same people were RB/RB heterozy-
RB- RB-
gotes. As Fig. A shows, mitotic recombination between the
RB gene and the centromere of the chromosome carrying Retinoblastoma
the gene provides one mechanism by which a cell in an
RB/RB individual could become RB/RB. Once a ho- Figure A How mitotic crossing-over can contribute to cancer.
mozygous RB cell is generated, it will divide uncontrol- Mitotic recombination during retinal growth in an RB/RB het-
lably, leading to tumor formation. erozygote may produce an RB/RB daughter cell that lacks a
Only 40% of retinoblastoma cases follow the preceding functional retinoblastoma gene and thus divides out of control.
scenario. The other 60% occur in people who are born with The crossover must occur between the RB gene and its cen-
two normal copies of the RB gene. In such people, it takes tromere. Only the arrangement of chromatids yielding this re-
two mutational events to cause the cancer. The first of these sult is shown.
Connections
Medical geneticists have used their understanding of link- the individuals in the pedigree were recombinant types. On
age, recombination, and mapping to make sense of the the other hand, even though hemophilia B is also on the X
pedigrees presented at the beginning of this chapter (see chromosome, it lies far enough away from the red-green
Fig. 5.1 on p. 124). The X-linked gene for red-green color- colorblindness locus that the two genes recombine rela-
blindness must lie very close to the gene for hemophilia A tively freely and may thus appear genetically unlinked, as
because the two are tightly coupled. In fact, an examination was the case in Fig. 5.1b. The recombination distance sep-
of many pedigrees has shown that if a doubly heterozygous arating the genes for colorblindness and hemophilia B is
woman receives defective alleles of both genes from the about 36 m.u. Pedigrees pointing to two different forms of
same parent, as few as 3% of her sons will have only one of hemophilia, one very closely linked to colorblindness, the
the conditions, as a result of recombination. This means other almost not linked at all, provided one of several indi-
that the genetic distance between the two genes is only 3 cations that hemophilia is determined by more than one
m.u. In Fig. 5.1a, the sample size was so small that none of gene (Fig. 5.26).
har48464_ch05_123-166 07/19/2006 08:53 PM Page 155
CONFIRMING PAGES
5.3 Mitotic Recombination Can Produce Genetic Mosaics 155
it is also possible to estimate the likelihood that the fetus

has received a closely linked disease-causing allele. DNA
markers have also been of great importance in the con-
struction of a physical map of the human genome, a
Hunter syndrome process we discuss in Chapters 10 and 11.
Hemophilia B The simultaneous occurrence of genetic linkage and
recombination is universal among life-forms and presum-
Fragile X syndrome ably confers some advantage to a species beyond allowing
geneticists to map its genes. Although definitive conclu-
Hemophilia A
sions are beyond our reach, speculation about the possible
G6PD deficiency: Favism advantages can help us understand why genomes are
Drug-sensitive anemia arranged as they are. Linkage provides the potential for
Chronic hemolytic anemia transmitting favorable combinations of genes intact to suc-
Colorblindness (several forms) cessive generations, while recombination produces great
Dyskeratosis congenita flexibility in generating new combinations of alleles. Some
Deafness with stapes fixation new combinations may help a species adapt to changing
TKCR syndrome environmental conditions, whereas the inheritance of suc-
Adrenoleukodystrophy cessfully tested combinations can preserve what has
Adrenomyeloneuropathy worked in the past. Linkage and recombination thus help
chromosomes function as organelles (in the broad sense)
Emery muscular dystrophy
that not only transmit but also evolve genetic information.
SED tarda
Spastic paraplegia, X-linked In Chapters 2–5 of this book, we have examined how
genes and chromosomes are transmitted. As important
Figure 5.26 A genetic map of part of the human and useful as this knowledge is, it tells us very little
X chromosome. about the structure and mode of action of the genetic ma-
terial. In the next section (Chapters 6–8), we carry our
analysis to the level of DNA, the actual molecule of
heredity. In Chapter 6, we look at DNA structure and
Refining the human chromosome map poses a contin- learn how the DNA molecule carries genetic information.
uous challenge for medical geneticists. The newfound po- In Chapter 7, we describe how geneticists defined the gene
tential for finding and fitting more and more DNA markers as a localized region of DNA containing many nucleotides
into the map (review the Fast Forward box in this chapter) that together encode the information to make a protein. In
enormously improves the ability to assign precise locations Chapter 8, we examine how the cellular machinery inter-
to genes that cause disease. If several markers flank one prets the genetic information in genes to produce the mul-
such gene and it is possible to test a fetus for those markers, titude of phenotypes that make up an organism.
Essential Concepts
1. Gene pairs that are close together on the same chro- Genes may assort independently either because they
mosome are genetically linked because they are are on different chromosomes or because they are far
transmitted together more often than not. The hall- apart on the same chromosome.
mark of linkage is that the number of parental types is
4. Statistical analysis helps determine whether or not
greater than the number of recombinant types among
two genes assort independently. The probability value
the progeny of double heterozygotes.
(p) calculated by the chi-square test measures the like-
2. The recombination frequencies of pairs of genes indi- lihood that a particular set of data supports the null
cate how often two genes are transmitted together. hypothesis of independent assortment, or no linkage.
For linked genes, the recombination frequency is less The lower the p value, the less likely is the null hy-
than 50%. pothesis, and the more likely the linkage. The chi-
square test can also be used to determine how well the
3. Gene pairs that assort independently exhibit a recom-
outcomes of crosses fit other genetic hypotheses (see
bination frequency of 50%, since the number of
www.mhhe.com/hartwell3: Chapter 3 for examples).
parental types equals the number of recombinants.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 156
CONFIRMING PAGES
5. The greater the physical distance between linked genes, 8. Organisms that retain all the products of one meiosis
the higher the recombination frequency. However, re- within an ascus reveal the relation between genetic
combination frequencies become more and more inac- recombination and the segregation of chromosomes
curate as the distance between genes increases. during the two meiotic divisions. Organisms like
Neurospora that produce ordered octads make it pos-
6. Recombination occurs because chromatids of homol-
sible to locate a chromosome’s centromere on the ge-
ogous chromosomes exchange parts (that is, cross
netic map.
over) during the prophase of meiosis I, after the chro-
mosomes have replicated. 9. In diploid organisms heterozygous for two alleles of
a gene, rare mitotic recombination between the gene
7. Genetic maps are a visual representation of relative
and its centromere can produce genetic mosaics in
recombination frequencies. The greater the density of
which some cells are homozygous for one allele or
genes on the map (and thus the smaller the distance
the other.
between the genes), the more accurate and useful the
map becomes in predicting inheritance.
On Our Website
www.mhhe.com/hartwell3
Chapter 5
Annotated Suggested Readings Specialized Topics
and Links to Other Websites
• The derivation and use of mapping functions
• The early history of genetic mapping
• Determining the linkage of human genes using likeli-
• Construction of a linkage map of the human genome hood ratios and LOD scores. A brief introduction to
this topic can also be found in the Tools of Genetics
• New ideas about the significance of chromosomal in-
box on pp. 416 – 417 of Chapter 11.
terference
• Using mitotic recombination to trace cells during de-
velopment
Social and Ethical Issues

1. In the initial stages of gene mapping, researchers find real risk of miscarriage when amniocentesis is per-
DNA markers that are linked to a disease gene (see formed to obtain fetal DNA, as explained in the Genet-
the Fast Forward box concerning cystic fibrosis on ics and Society box on p. 87 of Chapter 4.) Are most
pp. 142–143). It is possible to use such linked DNA patients intellectually and emotionally prepared to deal
markers to test for the presence of a disease-causing with a genetic diagnosis that is less than certain?
allele in a fetus, although these tests are by definition
2. In 1996, evidence was presented for linkage of a
inaccurate. For example, if the distance between the
marker to a predisposition to prostate cancer. Using 66
disease gene and the DNA marker is 5 m.u., then 5% of
families in which at least three males had prostate can-
fetuses diagnosed with the disease-causing allele would
cer, a genetic marker was found that is inherited with
actually be false positives, while 5% of the fetuses con-
high frequency in those men with prostate cancer.
sidered free of the disease-causing allele would be false
Linked to this marker, somewhere in a region of about
negatives and actually have the disease.
10 million base pairs, is a gene that predisposes to
How accurate must a genetic test be to be used on
prostate cancer. The estimate is that this genetic region
patients? To what extent does this decision depend
is involved in about 3% of the total cases of prostate
upon the severity of the genetic condition? How should
cancer. National newspapers announced these findings
one balance the costs and risks of the diagnostic proce-
with the following headlines:
dure? (For example, there is a very low but apparently
har48464_ch05_123-166 07/19/2006 08:53 PM Page 157
CONFIRMING PAGES
Solved Problems 157
“Scientists Find Proof That Mutant Gene Can Increase Risk use as a headline for an article on this study? What is the
of Prostate Cancer” responsibility of the press in presenting new findings?
“Prostate Cancer Gene Evidence Found”
Consider the responsibility of journalists/reporters to
their employers (newspapers), their profession, and
“Scientists Zero in on Gene Tied to Prostate Cancer” to the public.
Do these headlines accurately represent the situation?
Could they be misleading in any way? What would you
Solved Problems
I. The Xg locus on the human X chromosome has two al- 10% will be recombinant: sts a or Sts a (5% of
leles, a and a. The a allele causes the presence of each type). The phenotype of a son directly re-
the Xg surface antigen on red blood cells, while the reflects the genotype of the X chromosome from his
cessive a allele does not allow antigen to appear. The mother. Therefore, the probability that he will lack
Xg locus is 10 m.u. from the Sts locus. The Sts allele the Xg antigen and have icthyosis (genotype: sts
produces normal activity of the enzyme steroid sulfa- a / Y) is 45/100.
tase, while the recessive sts allele results in the lack of b. The probability that he will have the antigen and
steroid sulfatase activity and the disease ichthyosis ichthyosis (genotype: sts a / Y ) is 5/100.
(scaly skin). A man with ichthyosis and no Xg antigen c. There are two classes of gametes containing the
has a normal daughter with Xg antigen, who is expect- ichthyosis allele: sts a (45%) and sts a (5%). If
ing a child. the total number of gametes is 100, then 50 will
a. If the child is a son, what is the probability he will have the sts allele. Of those gametes, 5 (or 10%)
lack antigen and have ichthyosis? will have the a allele. Therefore there is a 1/10
b. What is the probability that a son would have both probability that a son with the sts allele will have
the antigen and ichthyosis? the Xg antigen.
c. If the child is a son with ichthyosis, what is the
probability he will have Xg antigen? II. Drosophila females of wild-type appearance but het-
erozygous for three autosomal genes are mated with
males showing three autosomal recessive traits: glassy
Answer
eyes, coal-colored bodies, and striped thoraxes. One
a. This problem requires an understanding of how thousand (1000) progeny of this cross are distributed
linkage affects the proportions of gametes. First in the following phenotypic classes:
designate the genotype of the individual in which
Wild type 27
recombination during meiosis affects the trans- Striped thorax 11
mission of alleles: in this problem, the daughter. Coal body 484
The X chromosome she inherited from her father Glassy eyes, coal body 8
(who had icthyosis and no Xg antigen) must be sts a. Glassy eyes, striped thorax 441
(No recombination could have separated the genes Glassy eyes, coal body, striped thorax 29
during meiosis in her father since he has only one a. Draw a genetic map based on this data.
X chromosome.) Because the daughter is normal b. Show the arrangement of alleles on the two
and has the Xg antigen, her other X chromosome homologous chromosomes in the parent females.
(inherited from her mother) must contain the Sts c. Normal-appearing males containing the same
and a alleles. Her X chromosomes can be chromosomes as the parent females in the pre-
diagrammed as: ceding cross are mated with females showing
glassy eyes, coal-colored bodies, and striped tho-
sts a
raxes. Of 1000 progeny produced, indicate the
numbers of the various phenotypic classes you
Sts a+
would expect.
Because the Sts and Xg loci are 10 m.u. apart on
the chromosome, there is a 10% recombination Answer
frequency. Ninety percent of the gametes will be A logical, methodical way to approach a three-point
parental: sts a or Sts a (45% of each type) and cross is described here.
har48464_ch05_123-166 07/19/2006 08:53 PM Page 158
CONFIRMING PAGES
a. Designate the alleles: The genetic map is

t wild-type thorax t striped thorax
c+ t+ g+
g wild-type eyes g glassy eyes
c wild-type body c coal-colored body
5.6 m.u. 1.9 m.u.
In solving a three-point cross, designate the types of
events that gave rise to each group of individuals and b. The alleles on each chromosome were already de-
the genotypes of the gametes obtained from their termined (c, g, t and c, g, t). Now that the or-
mother. (The paternal gametes contain only the reces- der of loci has also been determined, the
sive alleles of these genes [t g c]. They do not change arrangement of the alleles can be indicated.
the phenotype and can be ignored.) c t+ g+
Progeny Number Type of event Genotype
1. wild type 27 single crossover t g c c+ t g
2. striped thorax 11 single crossover t g c
3. coal body 484 parental t g c c. Males of the same genotype as the starting female
4. glassy eyes, 8 single crossover t g c (c t g / c t g) could produce only two types of
coal body gametes: parental types c t g and c t g because
5. glassy eyes, 441 parental t g c
striped thorax there is no recombination in male Drosophila. The
6. glassy eyes, 29 single crossover t g c progeny expected from the mating with a homozy-
coal body, gous recessive female are thus 500 coal body and
striped thorax 500 glassy eyes, striped thorax flies.
Picking out the parental classes is easy. If all the III. The following asci were obtained in Neurospora when
other classes are rare, the two most abundant cate- a wild-type strain (ad leu) was crossed to a double
gories are those gene combinations that have not un- mutant strain that cannot grow in the absence of ade-
dergone recombination. Then there should be two sets nine or leucine (ad leu). Only one member of each
of two phenotypes that correspond to a single spore pair produced by the final mitosis is shown, since
crossover event between the first and second genes, or the two cells in a pair have the same genotype. Total
between the second and third genes. Finally, there asci 120.
should be a pair of classes containing small numbers
Spore pair Ascus type
that result from double crossovers. In this example,

there are no flies in the double crossover classes, which 1–2 ad leu ad leu ad leu ad leu ad leu
would have been in the two missing phenotypic com- 3–4 ad leu ad leu ad leu ad leu ad leu
5–6 ad leu ad leu ad leu ad leu ad leu
binations: glassy eyes and coal body, striped thorax. 7–8 ad leu ad leu ad leu ad leu ad leu
Look at the most abundant classes to determine # of asci 30 30 40 2 18
which alleles were on each chromosome in the female
heterozygous parent. One parental class had the pheno- a. What genetic event causes the alleles of two genes
type of coal body (484 flies), so one chromosome in the to segregate to different cells at the second meiotic
female must have contained the t, g, and c alleles. division, and when does this event occur?
(Notice that we cannot yet say in what order these alle- b. Provide the best possible map for the two genes
les are located on the chromosome.) The other parental and their centromere(s).
class was glassy eyes and striped thorax, corresponding
to a chromosome with the t, g, and c alleles. Answer
To determine the order of the genes, compare the This problem requires an understanding of tetrad
tg c double crossover class (not seen in the data) analysis and the process (meiosis) that produces the
with the most similar parental class (t g c). The alle- patterns seen in ordered asci.
les of g and c retain their parental associations (g c), a. A crossover between a gene and its centromere
while the t gene has recombined with respect to both causes the segregation of alleles at the second
other genes in the double recombinant class. Thus, the meiotic division. The crossover event occurs dur-
t gene is between g and c. ing prophase of meiosis I.
In order to complete the map, calculate the recom- b. Using ordered tetrads you can determine whether
bination frequencies between the center gene and each two genes are linked, the distance between two
of the genes on the ends. For g and t, the nonparental genes, and the distance between each gene and its
combinations of alleles are in classes 2 and 4, so RF centromere. First designate the five classes of asci
(11 8) /1000 19 /1000, or 1.9%. For t and c, shown. The first class is a parental ditype (spores
classes 1 and 6 are nonparental, so RF (27 29)/ contain the same combinations of alleles as their
1000 56 / 1000, or 5.6%. parents); the second is a nonparental ditype; the
har48464_ch05_123-166 07/19/2006 08:53 PM Page 159
CONFIRMING PAGES
Problems 159
last three are tetratypes. Next determine if these its centromere is 35/2, or 17.5 m.u. Asci of the
genes are linked. The number of PD number of fourth and fifth types show a second-division seg-
NPD, so the genes are not linked. When genes are regation pattern for the ad gene
unlinked, the tetratype asci are generated by a
2 18
crossing-over event between a gene and its cen- 100 16.6%
tromere. Looking at the leu gene, there is a 120
second-division segregation pattern of that gene in Dividing 16.6% by 2 gives the recombination fre-
the third and fourth asci types. Therefore, the per- quency and map distance of 8.3 m.u. The map of
cent of second-division segregation is these two genes is the following:
40 2 ad
100 35%
120
8.3 m.u.
leu
Because only half of the chromatids in the meioses
that generated these tetratype asci were involved 17.5 m.u.
in the crossover, the map distance between leu and
Problems
Vocabulary the chi-square test to determine whether Mendel’s data
are significantly different from what he predicted. (The
1. Choose the phrase from the right column that best fits chi-square test did not exist in Mendel’s day, so he was
the term in the left column. not able to test his own data for goodness of fit to his
a. recombination 1. a statistical method for testing the fit
hypotheses.)
between observed and expected results 3. Two genes control color in corn snakes as follows:
b. linkage 2. an ascus containing spores of four O– B– snakes are brown, O– bb are orange, oo B– are
different genotypes
black, and oo bb are albino. An orange snake was mated
c. chi-square test 3. one crossover along a chromosome
makes a second nearby crossover
to a black snake, and a large number of F1 progeny
less likely were obtained, all of which were brown. When the F1
d. chiasma 4. when two loci recombine in less than snakes were mated to one another, they produced 100
50% of gametes brown offspring, 25 orange, 22 black, and 13 albino.
e. tetratype 5. the relative chromosomal location of a a. What are the genotypes of the F1 snakes?
gene b. What proportions of the different colors would
f. locus 6. the ratio of observed double crossovers have been expected among the F2 snakes if the two
to expected double crossovers loci assort independently?
g. coefficient of 7. individual composed of cells with c. Do the observed results differ significantly from
coincidence different genotypes
what was expected, assuming independent assort-
h. interference 8. formation of new genetic combinations
by exchange of parts between homologs
ment is occurring?
i. parental ditype 9. when the two alleles of a gene are
d. What is the probability that differences this great
segregated into different cells at the first between observed and expected values would hap-
meiotic division pen by chance?
j. ascospores 10. an ascus containing only two
nonrecombinant kinds of spores 4. A mouse from a true-breeding population with normal
k. first-division 11. structure formed at the spot where gait was crossed to a mouse displaying an odd gait called
segregation crossing-over occurs between homologs “dancing.” The F1 animals all showed normal gait.
l. mosaic 12. fungal spores contained in a sac a. If dancing is caused by homozygosity for the re-
cessive allele of a single gene, what proportion of
the F2 mice should be dancers?
Section 5.1
b. If mice must be homozygous for recessive alleles
2. Do the data that Mendel obtained fit his hypotheses? of both of two different genes to have the dancing
For example, Mendel obtained 315 yellow round, 101 phenotype, what proportion of the F2 should be
yellow wrinkled, 108 green round, and 32 green wrin- dancers if the two genes are unlinked?
kled seeds from the selfing of Yy Rr individuals (a total c. When the F2 mice were obtained, 42 normal and 8
of 556). His hypotheses of segregation and indepen- dancers were seen. Use the chi-square test to de-
dent assortment predict a 9:3:3:1 ratio in this case. Use termine if these results better fit the one-gene
har48464_ch05_123-166 07/19/2006 08:53 PM Page 160
CONFIRMING PAGES
model from part a or the two-gene model from cc dd parent; 903 Cc Dd, 897 cc dd, 98 Cc dd, and 102
part b. cc Dd offspring resulted.
a. How far apart are the c and d loci?
5. In Drosophila, males from a true-breeding stock with
b. What progeny and in what frequencies would you
raspberry-colored eyes were mated to females from a
expect to result from testcrossing the F1 genera-
true-breeding stock with sable-colored bodies. In the
tion from a CC dd cc DD cross to cc dd?
F1 generation, all the females had wild-type eye and
body color, while all the males had wild-type eye color 9. If the a and b loci are 20 m.u. apart in humans and an
but sable-colored bodies. When F1 males and females A B / a b woman mates with an a b / a b man, what is
were mated, the F2 generation was composed of 216 the probability that their first child will be A b / a b?
females with wild-type eyes and bodies, 223 females
10. In a particular human family, John and his mother both
with wild-type eyes and sable bodies, 191 males with
have brachydactyly (a rare autosomal dominant caus-
wild-type eyes and sable bodies, 188 males with rasp-
ing short fingers). John’s father has Huntington disease
berry eyes and wild-type bodies, 23 males with wild-
(another rare autosomal dominant). John’s wife is phe-
type eyes and bodies, and 27 males with raspberry eyes
notypically normal and is pregnant. Two-thirds of peo-
and sable bodies. Explain these results by diagram-
ple who inherit the Huntington (HD) allele show
ming the crosses, and calculate any relevant map dis-
symptoms by age 50, and John is 50 and has no symp-
tances.
toms. Brachydactyly is 90% penetrant.
6. Figure 5.5 on p. 130 applied the chi-square method to a. What are the genotypes of John’s parents?
test linkage between two genes by asking whether the b. What are the possible genotypes for John?
observed numbers of parental and recombinant classes c. What is the probability the child will express both
differed significantly from the expectation of inde- brachydactyly and Huntington disease by age 50 if
pendent assortment that parentals recombinants. the two genes are unlinked?
Another possible way to analyze the results from these d. If these two loci are 20 m.u. apart, how will it
same experiments is to ask whether the observed fre- change your answer to part c?
quencies of the four genotypic classes (A B, a b, A b,
11. In mice, the autosomal locus coding for the
-globin
and a B) can be explained by a null hypothesis predict-
chain of hemoglobin is 1 m.u. from the albino locus. As-
ing that they should appear in a 1:1:1:1 ratio. In order
sume for the moment that the same is true in humans.
to consider the relative advantages and disadvantages
The disease sickle-cell anemia is the result of homozy-
of analyzing the data in these two different ways an-
gosity for a particular mutation in the
-globin gene.
swer the following:
a. A son is born to an albino man and a woman with
a. What is the null hypothesis in each case?
sickle-cell anemia. What kinds of gametes will the
b. Which is a more sensitive test of linkage? (Ana-
son form, and in what proportions?
lyze the data in Fig. 5.5 by the second method.)
b. A daughter is born to a normal man and a woman
c. How would both methods respond to a situation in
who has both albinism and sickle-cell anemia.
which one allele of one of the genes causes re-
What kinds of gametes will the daughter form, and
duced viability?
in what proportions?
7. In mice, the dominant allele Gs of the X-linked gene c. If the son in part a grows up and marries the daugh-
Greasy produces shiny fur, while the recessive wild- ter in part b, what is the probability that a child of
type Gs allele determines normal fur. The dominant theirs will be an albino with sickle-cell anemia?
allele Bhd of the X-linked Broadhead gene causes
12. In corn, the allele A allows the deposition of antho-
skeletal abnormalities including broad heads and
cyanin (blue) pigment in the kernels (seeds), while aa
snouts, while the recessive wild-type Bhd allele yields
plants have yellow kernels. At a second gene, W–
normal skeletons. Female mice heterozygous for the
produces smooth kernels, while ww kernels are wrin-
two alleles of both genes were mated with wild-type
kled. A plant with blue smooth kernels was crossed to
males. Among 100 male progeny of this cross, 49 had
a plant with yellow wrinkled kernels. The progeny
shiny fur, 48 had skeletal abnormalities, 2 had shiny fur
consisted of 1447 blue smooth, 169 blue wrinkled, 186
and skeletal abnormalities, and 1 was wild type.
yellow smooth, and 1510 yellow wrinkled.
a. Diagram the cross described, and calculate the dis-
a. Are the a and w loci linked? If so, how far apart are
tance between the two genes.
they?
b. What would have been the results if you had
b. What was the genotype of the blue smooth parent?
counted 100 female progeny of the cross?
Include the chromosome arrangement of alleles.
8. CC DD and cc dd individuals were crossed to each c. If a plant grown from a blue wrinkled progeny
other, and the F1 generation was backcrossed to the seed is crossed to a plant grown from a yellow
har48464_ch05_123-166 07/19/2006 08:53 PM Page 161
CONFIRMING PAGES
Problems 161
smooth F1 seed, what kinds of kernels would be used as a marker to follow inheritance of the disease
expected, and in what proportions? allele. In an informative family (in which one parent
is heterozygous for both the disease allele and the
13. Albino rabbits (lacking pigment) are homozygous for
DNA marker in a known chromosomal arrangement
the recessive c allele (C allows pigment formation).
of alleles and his or her mate does not have the same
Rabbits homozygous for the recessive b allele make
alleles of the DNA variant), the reliability of such a
brown pigment, while those with at least one copy of B
marker as a predictor of the disease in a fetus is re-
make black pigment. True-breeding brown rabbits
lated to the map distance between the DNA marker
were crossed to albinos, which were BB. F1 rabbits,
and the gene causing the disease. Imagine that a man
which were all black, were crossed to the double re-
affected with the disease (genotype Dd) is heterozy-
cessive (bb cc). The progeny obtained were 34 black,
gous for the V1 and V2 forms of the DNA variant, with
66 brown, and 100 albino.
form V1 on the same chromosome as the D allele and
a. What phenotypic proportions would have been ex-
form V2 on the same chromosome as d. His wife is
pected if the b and c loci were unlinked?
V3V3 dd, where V3 is another allele of the DNA
b. How far apart are the two loci?
marker. Typing of the fetus by amniocentesis reveals
14. Write the number of different kinds of phenotypes, ex- that the fetus has the V2 and V3 variants of the DNA
cluding gender, you would see among a large number of marker. How likely is it that the fetus has inherited
progeny from an F1 mating between individuals of the disease allele D if the distance between the D
identical genotype that are heterozygous for one or two locus and the marker locus is (a) 0 m.u., (b) 1 m.u.,
genes (that is, Aa or Aa Bb) as indicated. No gene inter- (c) 5 m.u., (d) 10 m.u., (e) 50 m.u.?
actions means that the phenotype determined by one
17. In Drosophila, the recessive dp allele of the dumpy
gene is not influenced by the genotype of the other gene.
gene produces short, curved wings, while the recessive
a. One gene; A completely dominant to a.
allele bw of the brown gene causes brown eyes. In a
b. One gene; A and a codominant.
testcross using females heterozygous for both of these
c. One gene; A incompletely dominant to a.
genes, the following results were obtained:
d. Two unlinked genes; no gene interactions; A com-
pletely dominant to a, and B completely dominant
wild-type wings, wild-type eyes 178
to b. wild-type wings, brown eyes 185
e. Two genes, 10 m.u. apart; no gene interactions; A dumpy wings, wild-type eyes 172
completely dominant to a, and B completely dom- dumpy wings, brown eyes 181
inant to b.
f. Two unlinked genes; no gene interactions; A and a In a testcross using males heterozygous for both of
codominant, and B incompletely dominant to b. these genes, a different set of results was obtained:
g. Two genes, 10 m.u. apart; A completely dominant
to a, and B completely dominant to b; and with re- wild-type wings, wild-type eyes 247
cessive epistasis between the genes. dumpy wings, brown eyes 242
h. Two unlinked duplicated genes (that is, A and B
perform the same function); A and B completely a. What can you conclude from the first testcross?
dominant to a and b, respectively. b. What can you conclude from the second testcross?
i. Two genes, 0 m.u. apart; no gene interactions; A c. How can you reconcile the data shown in parts a
completely dominant to a, and B completely dom- and b? Can you exploit the difference between
inant to b. (There are two possible answers.) these two sets of data to devise a general test for
synteny in Drosophila?
15. If the a and b loci are 40 cM apart and an AA BB indi- d. The genetic distance between dumpy and brown is
vidual and an aa bb individual mate: 91.5 m.u. How could this value be measured?
a. What gametes will the F1 individuals produce,
and in what proportions? What phenotypic 18. Cinnabar eyes (cn) and reduced bristles (rd) are
classes in what proportions are expected in the F2 autosomal recessive characters in Drosophila. A ho-
generation (assuming complete dominance for mozygous wild-type female was crossed to a re-
both genes)? duced, cinnabar male, and the F1 males were then
b. If the original cross was AA bb aa BB, what ga- crossed to the F1 females to obtain the F2. Of the 400
metic proportions would emerge from the F1? F2 offspring obtained, 292 were wild type, 9 were
What would be the result in the F2 generation? cinnabar, 7 were reduced, and 92 were reduced,
cinnabar. Explain these results and estimate the dis-
16. A DNA variant has been found linked to a rare auto- tance between the cn and rd loci.
somal dominant disease in humans and can thus be
har48464_ch05_123-166 07/19/2006 08:53 PM Page 162
CONFIRMING PAGES
Section 5.2 21. In Drosophila, three autosomal genes have the follow-
ing map:
19. In Drosophila, the recessive allele mb of one gene
causes missing bristles, the recessive allele e of a sec- a b c
ond gene causes ebony body color, and the recessive
allele k of a third gene causes kidney-shaped eyes. 20 m.u. 10 m.u.
(Dominant wild-type alleles of all three genes are indi- a. Provide the data, in terms of the expected number
cated with a superscript.) The three different P gen- of flies in the following phenotypic classes, when
eration crosses in the table that follows were a b c / a b c females are crossed to a b c / a b c
conducted, and then the resultant F1 females from each males. Assume 1000 flies were counted and that
cross were testcrossed to males that were homozygous there is no interference in this region.
for the recessive alleles of both genes in question. The
phenotypes of the testcross offspring are tabulated as a b c
follows. Determine the best genetic map explaining all a b c
a b c
the data. a b c
a b c
Parental cross Testcross offspring of F1 females a b c

mb mb , e e
normal bristles, normal body 117 a b c
mb mb, e e normal bristles, ebony body 11 a b c
missing bristles, normal body 15
missing bristles, ebony body 107 b. If the cross were reversed, such that a b c / a b c
males are crossed to a b c / a b c females, how
k k, e e k k, e e normal eyes, normal body 11 many flies would you expect in the same pheno-
normal eyes, ebony body 150
kidney eyes, normal body 144 typic classes?
kidney eyes, ebony body 7 22. A snapdragon with pink petals, black anthers, and long
mb mb, k k normal bristles, normal eyes 203 stems was allowed to self-fertilize. From the resulting
mb mb, k k normal bristles, kidney eyes 11 seeds, 650 adult plants were obtained. The phenotypes
missing bristles, normal eyes 15 of these offspring are listed here.
missing bristles, kidney eyes 193
78 red long tan
26 red short tan
20. In the tubular flowers of foxgloves, wild-type col- 44 red long black
oration is red while a mutation called white produces 15 red short black
white flowers. Another mutation, called peloria, causes 39 pink long tan
the flowers at the apex of the stem to be huge. Yet an- 13 pink short tan
other mutation, called dwarf, affects stem length. You 204 pink long black
68 pink short black
cross a white-flowered plant (otherwise phenotypically 5 white long tan
wild type) to a plant that is dwarf and peloria but has 2 white short tan
wild-type red flower color. All of the F1 plants are tall 117 white long black
with white, normal-sized flowers. You cross an F1 plant 39 white short black
back to the dwarf and peloria parent, and you see the a. Using P for one allele and p for the other, indicate
543 progeny shown in the chart. (Only mutant traits how flower color is inherited.
are noted.) b. What numbers of red : pink : white would have
dwarf, peloria 172 been expected among these 650 plants?
white 162 c. How are anther color and stem length inherited?
dwarf, peloria, white 56 d. What was the genotype of the original plant?
wild type 48
dwarf, white 51
e. Do any of the three genes show independent
peloria 43 assortment?
dwarf 6 f. For any genes that are linked, indicate the arrange-
peloria, white 5 ments of the alleles on the homologous chromo-
somes in the original snapdragon, and estimate the
a. Which alleles are dominant?
distance between the genes.
b. What were the genotypes of the parents in the
original cross? 23. Male Drosophila expressing the recessive mutations sc
c. Draw a map showing the linkage relationships of (scute), ec (echinus), cv (crossveinless), and b (black)
these three loci. were crossed to phenotypically wild-type females, and
d. Is there interference? If so, calculate the coeffi- the 3288 progeny listed were obtained. (Only mutant
cient of coincidence and the interference value. traits are noted.)
har48464_ch05_123-166 07/19/2006 08:53 PM Page 163
CONFIRMING PAGES
Problems 163
653 black, scute, echinus, crossveinless d. What proportion of the backcross progeny will
670 scute, echinus, crossveinless have the leaf morphology and leaf color of the Vir-
675 wild type ginia strain but the leaf size of the Carolina strain?
655 black
71 black, scute 26. a. In Drosophila, crosses between F1 heterozygotes
73 scute of the form A b / a B always yield the same ratio of
73 black, echinus, crossveinless
74 echinus, crossveinless phenotypes in the F2 progeny regardless of the dis-
87 black, scute, echinus tance between the two genes (assuming complete
84 scute, echinus dominance for both autosomal genes). What is this
86 black, crossveinless ratio? Would this also be the case if the F1 het-
83 crossveinless erozygotes were A B / a b?
1 black, scute, crossveinless
1 scute, crossveinless b. If you intercrossed F1 heterozygotes of the form
1 black, echinus A b / a B in mice, the phenotypic ratio among the
1 echinus F2 progeny would vary with the map distance be-
tween the two genes. Is there a simple way to esti-
a. Diagram the genotype of the female parent.
mate the map distance based on the frequencies of
b. Map these loci.
the F2 phenotypes, assuming rates of recombina-
c. Is there evidence of interference? Justify your
tion are equal in males and females? Could you es-
answer with numbers.
timate map distances in the same way if the mouse
24. Drosophila females heterozygous for each of three re- F1 heterozygotes were A B / a b?
cessive autosome mutations with independent pheno-
27. The following list of four Drosophila mutations indi-
typic effects (thread antennae [th], hairy body [h], and
cates the symbol for the mutation, the name of the
scarlet eyes [st]) were testcrossed to males showing all
gene, and the mutant phenotype:
three mutant phenotypes. The 1000 progeny of this
testcross were Allele symbol Gene name Mutant phenotype
thread, hairy, scarlet 432 dwp dwarp small body, warped wings
wild type 429 rmp rumpled deranged bristles
thread, hairy 37 pld pallid pale wings
thread, scarlet 35 rv raven dark eyes and bodies
hairy 34
scarlet 33 You perform the following crosses with the indicated
a. Show the arrangement of alleles on the relevant results:
chromosomes in the triply heterozygous females.
Cross #1: dwarp, rumpled females pallid, raven males
b. Draw the best genetic map that explains these data.
→ dwarp, rumpled males and wild-type females
c. Calculate any relevant interference values.
25. A true-breeding strain of Virginia tobacco has domi- Cross #2: pallid, raven females dwarp, rumpled males
nant alleles determining leaf morphology (M), leaf → pallid, raven males and wild-type females
color (C), and leaf size (S). A Carolina strain is ho-
F1 females from cross #1 were crossed to males from a
mozygous for the recessive alleles of these three genes.
true-breeding dwarp rumpled pallid raven stock. The
These genes are found on the same chromosome as
1000 progeny obtained were as follows:
follows:
pallid 3
M C S pallid, raven 428
pallid, raven, rumpled 48
6 m.u. 17 m.u. pallid, rumpled 23
dwarp, raven 22
dwarp, raven, rumpled 2
An F1 hybrid between the two strains is now backcrossed dwarp, rumpled 427
to the Carolina strain. Assuming no interference: dwarp 47
a. What proportion of the backcross progeny will re-
semble the Virginia strain for all three traits? Indicate the best map for these four genes, including
b. What proportion of the backcross progeny will re- all relevant data. Calculate interference values where
semble the Carolina strain for all three traits? appropriate.
c. What proportion of the backcross progeny will 28. A cross was performed between one haploid strain of
have the leaf morphology and leaf size of the Vir- yeast with the genotype a f g and another haploid strain
ginia strain but the leaf color of the Carolina with the genotype (a and are mating types).
strain?
har48464_ch05_123-166 07/19/2006 08:53 PM Page 164
CONFIRMING PAGES
The resulting diploid was sporulated, and a random hundred asci were dissected, and colonies were grown
sample of 101 of the resulting haploid spores was ana- from the four spores in each ascus. Cells from these
lyzed. The following genotypic frequencies were seen: colonies were tested for their ability to grow on petri
plates containing either minimal medium (min), min
31
a f g 29 lysine (lys), min methionine (met), or min lys
a f 14 met. The asci could be divided into two groups based
g 13 on this analysis:
a g 6
f 6 Group 1: In 89 asci, cells from two of the four spore
a 1 colonies could grow on all four kinds of media, while
f g 1 the other two spore colonies could grow only on min
a. Map the loci involved in the cross. lys met.
b. Assuming all three genes are on the same chromo- Group 2: In 11 asci, cells from one of the four spore colonies
some arm, is it possible that a particular ascus could grow on all four kinds of petri plates. Cells from
could contain an f g spore but not an a a second one of the four spore colonies could grow
spore? If so, draw a meiosis that could generate only on min lys plates and on min lys met
such an ascus. plates. Cells from a third of the four spore colonies
29. Neurospora of genotype a c are crossed with Neu- could only grow on min met plates and on min lys
rospora of genotype b . The following tetrads are met. Cells from the remaining colony could only
obtained (note that the genotype of the four spore pairs grow on min lys met.
in an ascus are listed, rather than listing all eight
spores): a. What are the genotypes of each of the spores
within the two types of asci?
ac a b c c b c a b a c b. Are the lys and met genes linked? If so, what is the
ac a b c ac a b c a b a b c map distance between them?
b b c
b a b a c b
c. If you could extend this analysis to many more
asci, you would eventually find some asci with a
137 141 26 25 2 3
different pattern. For these asci, describe the phe-
a. In how many cells has meiosis occurred to yield notypes of the four spores. List these phenotypes
this data? as the ability of dissected spores to form colonies
b. Give the best genetic map to explain these results. on the four kinds of petri plates.
Indicate all relevant genetic distances, both be-
32. The a, b, and c loci are all on different chromosomes in
tween genes and between each gene and its re-
yeast. When a yeast were crossed to b yeast and
spective centromere.
the resultant tetrads analyzed, it was found that the
c. Diagram a meiosis that could give rise to one of
number of nonparental ditype tetrads was equal to the
the three tetrads in the class at the far right in the
number of parental ditypes, but there were no tetratype
list.
asci at all. On the other hand, many tetratype asci were
30. Two crosses were made in Neurospora involving the seen in the tetrads formed after a was crossed with
mating type locus and either the ad or p genes. In both c, and after b was crossed with c. Explain these
cases, the mating type locus (A or a) was one of the results.
loci whose segregation was scored. One cross was
33. Indicate the percentage of tetrads that would have 0,
ad A a (cross a), and the other was p A a
1, 2, 3, or 4 viable spores after Saccharomyces cere-
(cross b). From cross a, 10 parental ditype, 9 non-
visiae a / diploids of the following genotypes are
parental ditype, and 1 tetratype asci were seen. From
sporulated:
cross b, the results were 24 parental ditype, 3 non-
a. A true-breeding wild-type strain (with no muta-
parental ditype, and 27 tetratype asci.
tions in any gene essential for viability).
a. What are the linkage relationships between the
b. A strain heterozygous for a null (completely inac-
mating type locus and the other two loci?
tivating) mutation in a single essential gene.
b. Although these two crosses were performed in
Neurospora, you cannot use the data given to cal-
For the remaining parts of this problem, consider
culate centromere-to-gene distances for any of
crosses between yeast strains of the form a b, where
these genes. Why not?
a and b are both temperature-sensitive mutations in
31. A cross was performed between a yeast strain that re- different essential genes. The cross is conducted under
quires methionine and lysine for growth (met lys) permissive (low-temperature) conditions. Indicate the
and another yeast strain, which is met lys. One percentage of tetrads that would have 0, 1, 2, 3, or 4
har48464_ch05_123-166 07/19/2006 08:53 PM Page 165
CONFIRMING PAGES
Problems 165
viable spores subsequently measured under restrictive your answer to part d to obtain a more accurate
(high-temperature) conditions. equation for calculating map distances between
c. a and b are unlinked, and both are 0 m.u. from two genes from the results of tetrad analysis?
their respective centromeres. f. Using your corrected equation from part e, what is
d. a and b are unlinked; a is 0 m.u. from its cen- a more accurate measurement of the distance in
tromere, while b is 10 m.u. from its centromere. map units between the his and lys genes?
e. a and b are 0 m.u. apart.
f. a and b are 10 m.u. apart. Assume all crossovers 36. A research group has selected three independent trp
between a and b are SCOs (single crossovers). haploid strains of Neurospora, each of which cannot
g. In part (f), if a four-strand DCO (double crossover) grow in the absence of the amino acid tryptophan.
occurred between a and b, how many of the spores They first mated these three strains with a wild-type
in the resulting tetrad would be viable at high tem- strain of opposite mating type, and then they analyzed
perature? the resultant octads. For all three matings, two of the
four spore pairs in every octad could grow on minimal
34. Two genes are located on the same chromosome as medium (that is, in the absence of tryptophan), while
follows: the other two spore pairs were unable to grow on this
minimal medium.
c d
a. What can you conclude from this result?
7 m.u. 15 m.u. In the matings of mutant strains 1 and 2 with wild type,
one of the two topmost pairs in some octads had spores
A haploid cross of the form C D c d is made. that could grow on minimal medium while the other of
a. What proportions of PD, NPD, and T tetrads the two topmost pairs in the same octads had spores
would you expect if this cross was made between that could not grow on minimal medium. In the mating
strains of Saccharomyces cerevisiae and the inter- of mutant strain 3 with wild type, either all the spores
ference in this region 1? in the two topmost pairs could grow on minimal
b. If the interference in this region 0? medium or all could not grow on minimal medium.
c. What kinds of tetrads, and in what proportions, b. What can you conclude from this result?
would you expect if this cross was made between The researchers next prepared two separate cultures of
strains of Neurospora crassa and the interference in each mutant strain; one of these cultures was of mating
this region 1? (Consider not only whether a tetrad type A and the other of mating type a. They mated these
is PD, NPD, or T but also whether the tetrad shows strains in pairwise fashion, dissected the resultant octads,
first or second division segregation for each gene.) and determined how many of the individual spores could
d. If the interference in this region 0? grow on minimal medium. The results are shown here.
35. A yeast strain that cannot grow in the absence of the
amino acid histidine (his) is mated with a yeast strain % of octads with x number of spores viable on
Mating minimal medium
that cannot grow in the absence of the amino acid
lysine (lys). Among the 400 unordered tetrads result- x0 2 4 6 8
ing from this mating, 233 were PD, 11 were NPD, and 1 2 78 22 0 0 0
156 were T. 1 3 46 6 48 0 0
a. What types of spores are in the PD, NPD, and T 2 3 42 16 42 0 0
tetrads?
b. What is the distance in map units between the his c. For each of the three matings in the table, how
and lys genes? many of the 100 octads are PD? NPD? T?
c. Assuming that none of these tetrads was caused by d. Draw a genetic map explaining all of the preced-
more than two crossovers between the genes, how ing data. Assume that the sample sizes are suffi-
can you estimate the number of meioses that gen- ciently small that none of the octads are the result
erated these 400 tetrads in which zero, one, or two of double crossovers.
crossovers took place? e. Although this problem describes crosses in
d. Based on your answer to part c, what is the mean Neurospora, it does not help in this particular case
number of crossovers per meiosis in the region be- to present the matings in the table as ordered
tween the two genes? octads. Why not?
e. The equation RF 100 (NPD 1/2T) / total f. Why in this particular problem can you obtain
tetrads accounts for some, but not all, double gene–centromere distances from the crosses in the
crossovers between two genes. Which double table, even though the data are not presented as or-
crossovers are missed? Can you extrapolate from dered octads?
har48464_ch05_123-166 07/19/2006 08:53 PM Page 166
CONFIRMING PAGES
Section 5.3 40. Neurofibromas are tumors of the skin that can arise
when a skin cell that is originally NF1 / NF1 loses
37. A single yeast cell placed on a solid agar will divide
the NF1 allele. This wild-type allele encodes a func-
mitotically to produce a colony of about 107 cells. A tional tumor suppressor protein, while the NF1 allele
haploid yeast cell that has a mutation in the ade2 gene encodes a nonfunctional protein.
will produce a red colony; an ade2 colony will be A patient of genotype NF1 / NF1 has 20 inde-
white. Some of the colonies formed from diploid yeast pendent tumors in different areas of the skin. Samples
cells with a genotype of ade2 / ade2 will contain are taken of normal, noncancerous cells from this pa-
sectors of red within a white colony.
tient, as well as of cells from each of the 20 tumors.
a. How would you explain these sectors?
Extracts of these samples are analyzed by a technique
b. Although the white colonies are roughly the same
called gel electrophoresis that can detect variant forms
size, the red sectors within some of the white
of four different proteins (A, B, C, and D) all encoded
colonies vary markedly in size. Why? Do you ex-
by genes that lie on the same autosome as NF1. Each
pect the majority of the red sectors to be relatively
protein has a slow (S) and a fast (F) form that are en-
large or relatively small?
coded by different alleles (for example, AS and AF). In
38. A diploid strain of yeast has a wild-type phenotype but the extract of normal tissue, slow and fast variants of
the following genotype: all four proteins are found. In the extracts of the tu-
mors, 12 had only the fast variants of proteins A and D
but both the fast and slow variants of proteins B and C;
b a c leth d e
6 had only the fast variant of protein A but both the fast
and slow variants of proteins B, C, and D; and the re-
b+ a+ c+ leth + d+ e+ maining 2 tumor extracts had only the fast variant of
protein A, only the slow variant of protein B, the fast
and slow variants of protein C, and only the fast variant
a, b, c, d, and e all represent recessive alleles that yield
of protein D.
a visible phenotype, and leth represents a recessive
a. What kind of genetic event described in this chap-
lethal mutation. All genes are on the same chromo-
ter could cause all 20 tumors, assuming that all the
some, and a is very tightly linked to its centromere
tumors are produced by the same mechanism?
(indicated by a small circle). Which of the following
b. Draw a genetic map describing these data, assum-
phenotypes could be found in sectors resulting from
ing that this small sample represents all the types of
mitotic recombination in this cell? (1) a; (2) b; (3) c;
tumors that could be formed by the same mecha-
(4) d; (5) e; (6) b e; (7) c d; (8) c d e; (9) d e; (10) a b.
nism in this patient. Show which alleles of which
Assume that double mitotic crossovers are too rare to
genes lie on the two homologous chromosomes.
be observed.
Indicate all relative distances that can be estimated.
39. In Drosophila, the yellow (y) gene is near the end of c. Another mechanism that can lead to neurofibro-
the acrocentric X chromosome, while the singed (sn) mas in this patient is a mitotic error producing
gene is located near the middle of the X chromosome. cells with 45 rather than the normal 46 chromo-
On the wings of female flies of genotype y sn / y sn, somes. How can this mechanism cause tumors?
you can very rarely find patches of yellow tissue within How do you know, just from the results described,
which a small subset of cells also have singed bristles. that none of these 20 tumors is formed by such mi-
a. How can you explain this phenomenon? totic errors?
b. Would you find similar patches on the wings of d. Can you think of any other type of error that could
females having the genotype y sn / y sn? produce the results described?

Linkage and Recombination PDF

Uploaded by

Copyright:

Available Formats

Linkage and Recombination PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Linkage and Recombination PDF

Uploaded by

Copyright:

Available Formats

har48464_ch05_123-166 07/19/2006 08:52 PM Page 123

PART I Basic Principles: How Traits Are Transmitted

124 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

Our presentation of the transmission patterns of genes on the same chromosome

genes that do not encode proteins. Recognition that many

5.1 Gene Linkage and Recombination 125

multigenerational crosses yielding large numbers of P

126 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

Cross Series A Cross Series B

Parental Parental Recombinant Recombinant Parental Parental Recombinant Recombinant

5.1 Gene Linkage and Recombination 127

Percentages of Parental and Recombinant P b c +/ b c + b+ c / b+ c

derance of parental combinations among the F2 genotypes Total 7419

experiment received parental gene combinations (that

128 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

The Chi-Square Test

5.1 Gene Linkage and Recombination 129

TABLE 5.1 Critical Chi-Square Values

Cannot Reject the Null Hypothesis Null Hypothesis Rejected

2 0.02 0.21 1.39 4.61 5.99 9.21 13.82

3 0.11 0.58 2.37 6.25 7.81 11.35 16.27

5 0.55 1.61 4.35 9.24 11.07 15.09 20.52

130 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

5.1 Gene Linkage and Recombination 131

132 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

Additional material from

car+ Bar+ car+ Bar+

car Bar car Bar+

Chromosomes Parental Recombinant

5.1 Gene Linkage and Recombination 133

(a) Duplicated chromosome homologs (a) w y

4 Figure 5.8 Recombination frequencies are the basis of

134 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

5.2 Mapping; Locating Genes Along a Chromosome 135

Along a Chromosome 1.1+ 32.1 + 4.0 + 17.8 = 5

Figure 5.9 Mapping genes by comparisons of two-point

136 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

m would be located in the vicinity of 34 m.u. But what about

ble to obtain enough information to position the three 4197

252  241  13  9 Figure 5.11 depicts the homologous autosomes of the F1

5.2 Mapping: Locating Genes Along a Chromosome 137

(b) Crossover in region 1 Resultant chromatids

(c) Crossover in region 2 Resultant chromatids

(d) Double crossover; one crossover in each region Resultant chromatids

138 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

5.2 Mapping: Locating Genes Along a Chromosome 139

experimental progeny because one exchange effectively

140 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

5.2 Mapping: Locating Genes Along a Chromosome 141

33.0 vermilion eyes

44.0 garnet eyes 48.5 black body 44.0 scarlet eyes

91.1 rough eyes

100.5 plexus wings 100.7 claret eyes

142 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

Gene Mapping May Lead to a Cure for Cystic Fibrosis

5.2 Mapping: Locating Genes Along a Chromosome 143

By 1986, linkage analyses of hundreds of DNA markers Chromosome 7

• Finally, investigators discovered linkage with several

144 Chapter 5 Linkage, Recombination, and the Mapping of Genes on Chromosomes

252 241 13 9 Figure 5.11 depicts the homologous autosomes of the F1