Principles of Evolution Web Page
Lectures
LECTURE 18: TESTING EVOLUTIONARY HYPOTHESES
USING PHYLOGENIES
I. Introduction
A. In the last lecture we learned how to reconstruct
phylogenies.
B. In this lecture, we will examine some examples
of how reconstructed phylogenies can be used to
make inferences about historical
evolutionary processes.
C. Before doing so, however, we will first need
to discuss briefly how character states are mapped
onto phylogenies.
II. Mapping character states onto phylogenies
A. Mapping character states onto phylogenies essentially
means assigning character states to
the nodes of a phylogeny--Note
that this is a different problem from reconstructing the
phylogeny itself.
B. Methods
1. Both parsimony criteria and
likelihood techniques can be used to map character states.
2. We will discuss a parsimony
method because of its simplicity
C. Character state reconstruction using parsimony:
the uppass-downpass alogarithm
1. Start with a phylogeny, with
the character state known for each terminal taxon:
2. Start with two terminal taxa
that share an immediate common ancestor (e.g. Taxa V and VI
in the figure).
3. Assign character state(s)
to the common ancestor of those two taxa using two rules:
a. If the
two taxa share any character state(s) in common, assign the set of shared
states to the common ancestor.
b. If the
two taxa share no character states in common, assign the union of character
state(s) exhibited by each of the taxa to the common ancestor.
4. For example, node F, the common
ancestor of Taxa V and VI, would be assigned a state of
0,1 (i.e.,
can't tell whether it's 0 or 1), since the state of Taxon V is 1, while the
state of
Taxon VI is
0 (Rule b.)
5. Repeat 3. for any other two
terminal taxa sharing a common ancestor.
a. For example,
node D, the common ancestor of Taxa I and II, is assigned a state
of 0, 1. (Rule b.)
6. Procede down through the tree,
assigning successive ancestral nodes based on previous
assignments
of corresponding descendent nodes.
a. For example,
node D is assigned a state of 1, since it's immediate descendents are
Taxon IV (state 1) and node E (state 0, 1) . (Rule a)
7. This "downpass" results in
the following assignments to the nodes of the tree:
8. The next step is an "uppass"
a. Start with
the most ancestral node and compare it individually to each descendent
internal node, applying rules (a) and (b) to determine the state of the descendent
node.
b. For example,
compare nodes A and C. Because A and C share state 0, Rule (a) says
we should assign state 0 to node C.
c. Then compare
node C (state=0) to node E (state= 1). Since no states are shared,
Rule
(b) says assign state 0, 1 to node E.
d. Then compare
node E (state= 0, 1) to node F (state = 0, 1). According to Rule (a),
assign state 0, 1 to node F.
e. Finally,
compare node B (state=0) to node D (state= 0, 1). According to Rule
(a),
assign state 0 to node D.
9. These uppass assignments yield
the final optimized character states:
10. Note that some character
states are ambiguous (e.g. nodes E and F), while others are
completely resolved (e.g. nodes A, B, C, and D).
11. The ambiguities correspond
to two equally parsimonious trees:
III. Testing Evolutionary Hypotheses using Phylogenies: Evolution of
aggregation in butterflies:
B. Sillen-Tullberg. 1988. Evolution of gregariousness
in aposematic butterflies: a phylogenetic
analysis. Evolution 42: 293-305.
A. Natural History
1. Investigations of the basic
biology of butterflies have revealed a correlation between two
traits of
butterfly larvae: gregariousness and aposematism.
2. In most species of butterflies,
larvae are solitary. However, in some species, females lay
eggs in clusters
and the larvae remain together as a group as they feed and grow.
3. The larvae of most butterfly
species are cryptic, which protects them from discovery
by predators.
However, in some species, larvae exhibit aposematic coloration: they are
brightly colored
and conspicuous to predators. They are also distasteful.
The
conspicuous
coloration advertises this distastefulness to predators, which avoid
feeding on
the larvae.
4. In a survey of butterfly species,
it is generally found that species that are aposematic
are more frequently
gregarious than would be expected by chance association.
B. Hypotheses to explain this correlation
1. There have been two general
hypotheses put forward to explain this correlation, one involving
kin seleciton
and one involving individual selection.
2. Kin selection argument:
a. To explain:
how aposematism can evolve.
b. This argument
assumes that being conspicuously colored is always individually
disadvantageous because it calls the attention of the predator to the lava.
The
predator attacks the larva and kills it, and then learns, for a while, to
associate
the conspicuous coloration with distastefulness.
c. When
larvae are solitary, a mutation that arises causing conspicuousness provides
no benefit to other individuals because any other siblings that carry the
same
mutation and are conspicuous occur in other places and are thus not likely
to be
encountered by the predator before it "forgets" the association between aposematism
and distastefulness
d. There is
thus no altruistic advantage directed toward close relative that could
offset the individual disadvantage of aposematism, and natural selection
is thus
expected to eliminate the alleles causing it.
e. By contrast,
when larvae occur together in sib groups, other individuals in the
group are likely also to carry the mutation conferring conspicuous coloration.
f. When
the predator samples one conspicuousness individual and learns to associate
aposematism with distastefulness, this protects the other aposematically
colored
siblings in the group, thus reducing their probability of mortality.
g. This altruistic
benefit enjoyed by the kin of the sampled larvae may be sufficient to
overcome the individual disadvantage and favor the evolution of aposematism.
h. Thus, according
to this argument, aposematism will tend to evolve in gregarious
species but not in solitary species.
i. This
argument also makes a prediction about the order in which aposematism and
gregariousness evolve: gregariousness evolves first, since it is a
prerequisite to
kin selection favoring the evolution of aposematism.
3. Individual selection argument
a. This argument
assumes that a predator can taste a larva and reject it without killing
it.
b. If this
is possible and frequent, then it is advantageous to an INDIVIDUAL to be
distasteful and to advertise that distastefulness by being conspicuous, since
conspicuousness will help a predator remember it has already sampled the
distasteful
individual.
c. Finally,
once individuals have evolved to be distasteful and aposematic, it is
advantageous for an individual to associate with other individuals so that
any
accidental mortality caused by predator sampling is likely to fall on the
another
individual.
d. Thus, this
argument accounts for the evolution of gregariousness and aposematism
by assuming that each of these traits is advantageous to the individual that
possesses them.
e. According
to this argument, gregariousness will tend to evolve in aposematic species
but not in cryptic species.
f. This
argument also makes a prediction about the order in which aposematism and
gregariousness evolve: aposematism evolves first, followed by gregariousness.
C. Distinguishing between the hypotheses
1. The two hypotheses can thus
be distinguished by the order in which they predict aposematism
and gregariousness
evolve.
2. By reconstructing butterfly
phylogeny and mapping gregariousness and aposematism onto
the phylogeny,
one can ascertain which character has tended to evolve first, and thus
distinguish
between the two hypotheses.
3. In particular, there are three
orderings that a phylogeny can reveal:
a. In tree
I in the figure, transitions to gregariousness and aposematism map
to the same location, so that no information is provided for distinguishing
between the hypotheses.
b. In tree
II, the transition to aposematism maps earler than the transition to
gregariousness, a result that would support the individual selection hypothesis.
c. In tree
III, the transition to gregariousness maps earlier than the transition
to aposematism, a result that would support the kin selection hypothesis.
D. Results
1. Sillen-Tullberg examined 10
phylogenies, each corresponding to a different group of
butterflies,
mapping changes in character states onto the phylogenies.
2. Example: the tribe Papilionini
in the family Papilionidae (swallowtails). The phylogeny
shows two
independent cases of the evolution of aposematism (W on the phylogeny,
for "warning"
coloration). One of these was followed by the evolution of gregariousness
(G in the
pylogeny), corresponding to tree type II in the above figure. The second
was
not followed
or preceded by the evolution of gregariousness.
3. Example: the european
species of the tribe Pierinae in the family Pieridae (sulphurs, etc.).
The phylogeny
shows four independent transitions to aposematism (W).
a. In two
of these cases, a transition to gregariousness (G) occurred at a time
indistinguishable from the transition to aposematism, as in tree type I.
These
cases thus provide no information helpful for distinguishing the hypotheses.
b. In two
cases, the evolution of aposematism occurred in solitary species, indicating
the evolution of gregariousness is not a precondition for the evolution of
aposematism.
4. In total, in the 10 phylogenies
Sillen-Tullberg examined there were
a. 3 cases
in which the ordering of the evolution of gregariousness and aposematism
was ambiguous, as in tree type I above.
b. 15 cases
in which the evolution of gregariousness followed the
evolution of aposematism, as in tree type III.
c. No cases
in which the evolution of gregariousness preceded the evolution of
aposematism, as in tree type II.
d. In addition,
there were 9 cases in which the evolution of aposematism occured in
lineages with solitary larvae, indicating that gregariousness is not necessary
for the
evolution of aposematism.
e. By contrast,
there were no cases in which the evolution of gregariousness occurred
in lineages with cryptic larvae, suggesting that aposematism may be necessary
for
the evolution of gregariousness.
5. Conclusion: Mapping
of character-state transitions onto butterfly phylogenies strongly
supports the individual advantage hypothesis over the kin selection
hypothesis as an explanation for the evolution of aposematism and
gregariousness.
IV. Testing Evolutionary Hypotheses using Phylogenies: Evolution of
Color Vision
(Yokoyama, et al. 1993.
Paralogous origin of the red- and green-sensitive visual pigment genes in
vertebrates. Mol.
Biol. Evol. 10: 527-538)
A. Natural History of Color Pigment Genes
1. In vertebrates, there is a small
family of genes coding for visual pigment proteins.
2. One pigment, sensitive primarily
to blue light, is located on an autosome.
3. All vertebrates also carry
another pigment gene, which is sensitive
to either
red or green.
4. Species with just these two
genes have dichromatic vision.
5. In some species, however,
including some fish, some reptiles, most birds and some primates
have an additional
pigment gene, which confers trichromatic vision
(three visual
pigments, for blue, red and green).
Distribution of dichromatic (D) and trichromatic (T)
color vision in vertebrates.
6. It is clear from the sequence
similarity between the red and green pigment genes on the
X chromosome,
as well as their proximity on the chromosome, that one gene has arisen
via
duplication from the other gene.
B. Two evolutionary hypotheses to explain the distribution
of trichromatic visual system in vertebrates
1. Hypothesis 1: There was a
unique duplication event that facilitated the evolution of
trichromatic
vision in a common ancestor of all vertebrates, with subsequent loss of
trichromatism
in many vertebrate lines.
a. Under this hypothesis,
the groups exhibiting trichromatic vision retain an ancestral
character.
2. Hypothesis 2: There
were repeated, independent duplication events that facilitated separate
evolution
of trichromatism in different groups.
a. Under this
hypothesis, the ancestral character was dichromatic vision, and
b. different
groups exhibiting trichromatic vision represent convergent evolution.
C. Distinguishing between these hypotheses
1. Procedure: sequence Red and
Green color pigment genes from fish (astyanax fasciatus) and
humans and
construct a phylogeny.
2. Expectation under Hypothesis
1: a single gene duplication occurred before speciation. The
resulting
gene tree is represented in the next figure.
a. Left figure
portrays the relative timing of gene duplication and speciation on the
gene lineages.
b. Right figure
portrays the resulting phylogeny.
c. Under this
hypothesis, the Red genes from fish and man are more closely related to
each other than either are to either Green gene, and vice versa.
3. Expectation under Hypothesis
2: two gene duplications occurred after speciation, one in the
fish lineage
and one in the human lineage. The resulting gene tree isrepresented
in the next
figure.
a. Figures as previously
b. Under this
hypothesis, the Red and Green genes from fish are more closely related
than either are to any human gene, and vice veras.
4. The phylogenetic analysis
of the sequences indicates that the second gene tree is correct.
Phylogenetic tree
of visual pigment genes. R: red-sensitive pigment; G: green-sensitive
pigment. Subscripts as follows: Hs--humans; Gg--chickens; Gge--gecko;
Af--fish. Note that the fish Astyanax fasciatus has two copies
of the green-sensitive gene. (From Yokoyama et al. 1993.
Mol. Biol. Evol. 10: 527-538)
5. Consequently, trichromatic
vision in fish and humans represents convergent evolution rather
than a commonly
inherited ancestral trait.
D. Adaptive changes in pigment genes.
1. There are three convergent
amino acid substitutions separating Red and Green pigment genes
that
occur in both the fish lineage and the human lineage:
amino acid position
Red Gene Green Gene
180
Ser
Ala
270
Tyr
Phe
285
Thr
Ala
2. Functional studies show that
these amino acid positions the wavelength of light to which
the visual
pigment is most sensitive.
[Lecture Outlines | Bio 120 Home Page | Department of Biology | Duke University ]
