Phylogeny: Reconstructing Evolutionary Trees (Part 2)

1
PhylogenyReconstructing Evolutionary Trees
(Part 2)

• Chapter 14

2
Cladistic vs. phenetic trees

• Cladistic trees are built from shared derived
  characters (synapomorphies), which are assumed to
  be homologous unless there is good reason (deeper
  analysis, parsimony) to believe otherwise
• requires polarizing characters or character
  states
• only synapomorphies are informative
• methodology is explicitly phylogenetic
• Phenetic trees are based on overall similarity
• Phenetic trees are phylogenetic only to the
  extent that degree of similarity closeness of
  relationship

3
An example to show how cladistic and phenetic
approaches can result in different trees 1

• Four taxa (W, X, Y, Z) and five characters (1
  5), each of which has two states

Characters Characters Characters Characters Characters
Taxon 1 2 3 4 5
W a b c d e
X a b c d e
Y a b c d e
Z a b c d e
4
An example to show how cladistic and phenetic
approaches can result in different trees 2

• Matrix of shared character states (similarity
  matrix)

Number of similar character states Number of similar character states Number of similar character states Number of similar character states
Taxon W X Y Z
W 2 4 3
X 1 0
Y 4
Z

• Start tree by pairing taxa W and Y, then assess
  similarities between W/Y, X, and Z (could also
  have started with Y/Z pair)

5
An example to show how cladistic and phenetic
approaches can result in different trees 3

• Three taxa (W/Y, X, Z) and four characters (1
  5), each of which has two states (discard
  character 2 because W and Y do not share the same
  state)

Characters Characters Characters Characters Characters
Taxon 1 2 3 4 5
W/Y a c d e
X a c d e
Z a c d e
6
An example to show how cladistic and phenetic
approaches can result in different trees 4

• Start tree by pairing taxa W and Y, then assess
  similarities between W/Y, X, and Z

Number of similar character states Number of similar character states Number of similar character states
Taxon W/Y X Z
W/Y 1 3
X 0
Z
In step 2, we pair W/Y with Z, and we are finished
7
Phenogram 1 for taxa W, X, Y, and Z
Decresasing similarity
8
Phenogram 2 for taxa W, X, Y, and Z
Decresasing similarity
9
Phenogram 3 for taxa W, X, Y, and Z
Decresasing similarity
10
Phenogram summary

• Our phenetic analysis strongly indicates that W,
  Y, and Z form a similar group that is quite
  different from X
• If we use this as an estimate of phylogeny, we
  conclude that W, Y, and Z are more closely
  related to one another than either is to X
• This is essentially the Unweighted Pair-Group
  Method (UPGMA)

11
A cladistic analysis of the same data

• The character states are now polarized (perhaps
  by comparison to an outgroup), such that the
  character states with asterisks () represent the
  derived character states

12
The character state matrix

• Four taxa (W, X, Y, Z) and five characters (1
  5), each of which has two states ( derived
  state)

Characters Characters Characters Characters Characters
Taxon 1 2 3 4 5
W a b c d e
X a b c d e
Y a b c d e
Z a b c d e
13
The character state matrix

• Four taxa (W, X, Y, Z) and five characters (1
  5), each of which has two states ( derived
  state)

Characters Characters Characters Characters Characters
Taxon 1 2 3 4 5
W a b c d e
X a b c d e
Y a b c d e
Z a b c d e
14
The character state matrix

• Four taxa (W, X, Y, Z) and five characters (1
  5), each of which has two states ( derived
  state)

Characters Characters Characters Characters Characters
Taxon 1 2 3 4 5
W a b c d e
X a b c d e
Y a b c d e
Z a b c d e
15
Cladogram for taxa W, X, Y, and Z
W
Y
Z
X
c,d,e
b
a
a,b,c,d,e
16
Phenogram 3 for taxa W, X, Y, and Z
Decresasing similarity
17
Cladogram summary

• There is only one most parsimonious cladogram
• Only characters 1 and 2 are informative
  (synapomorphies)
• W and X are sister taxa
• Z is the most distantly related of the four taxa,
  rather than closely related to W and Y
• The reason that X appears unrelated to the other
  three taxa in the phenogram is because X has
  three autapomorphies that make it dissimilar to
  the other three taxa.

18
Phylogeny and Classification 1

• Linnaean classification is based historically on
  morphological traits it is a phenetic
  classification system
• Species are defined by type specimens
• similar species are grouped into a genus
• genera, families, orders, classes, phyla,
  kingdoms
• Linnaeus lived a century before Darwin he was
  not an evolutionist and did not believe that his
  classification system described evolutionary
  relationship
• Darwin, however, recognized that the ability to
  construct a hierarchical classification system
  based on similarity is exactly what would be
  expected under his concept of evolutionary
  history as a tree that described descent from
  nested sets of common ancestors

19
Phylogeny and Classification 2

• Should classification reflect phylogeny?
• If our phylogeny of the whales and artiodactyls
  is correct, then whales are just a subgroup of
  the Order Artiodactyla, not an order of their own
  (Cetacea)
• Recognizing Cetacea as a separate order on a par
  with artiodactyls makes Artiodactyla a
  paraphyletic taxon a taxon that does not
  include all the descendants of its common ancestor

20
Phylogeny of whales and artiodactyls based on
presence/absence of SINEs and LINEs (Nikaido et
al. 1999) (Fig. 14.8)
Red line encloses traditional artiodactyl species
21
A monophyletic group all the decendants of a
common ancestor ( the common ancestor) (Fig.
14.10a)
22
A paraphyletic group does not include all the
descendants of the common ancestor (Fig. 14.10b)
23
Examples of paraphyletic taxa (Fig. 14.10c)
24
A polyphyletic group does not include the
common ancestor (Fig. 14.10b)
25
How to classify?

• A strict cladistic classification scheme would
  require a taxonomic level for every level of
  branching in a phylogeny might be extremely
  unwieldy
• Just about everyone would probably agree that
  polyphyletic taxa should be avoided (suggests
  non-existent evolutionary relationship)
• Evolutionary classification
• Recognizes grade as well as clade as a basis
  for classification
• Cetacea are sufficiently different (adaptations
  for fully aquatic existence) from other mammals
  that they should be given the status of an order,
  equivalent to the other orders of mammals
  (Primates, Carnivora, Rodentia, Artiodactyla,
  etc.)
• Paraphyletic taxa are justified when a great deal
  of morphological/physiological change occurs
  along one branch of a clade

26
Reptilia as a paraphyletic taxon

• Virtually all cladistic analyses of birds and
  reptiles agree that crocodilia and birds are
  sister groups that is, crocodiles are more
  closely related to birds than to other
  conventional reptiles such as snakes and lizards
• Putting birds in the class Aves makes the class
  Reptilia paraphyletic
• The justification is that birds (warmblooded,
  feathers, flight) seem to have attained a
  different grade than reptiles (cold blooded, no
  feathers)

27
Phylogeny of the main vertebrate groupsreptiles
are a paraphyletic group, made up of turtles,
lizards, snakes, and crocodiles
28
Phylogeny of the main vertebrate groupsreptiles
are a paraphyletic group, made up of turtles,
lizards, snakes, and crocodiles
29
Bootstrapping Trees 1

• How much confidence do we have in any particular
  tree?
• How dependent is it on the particular set of
  characters that we have analyzed?
• Would we have obtained the same tree if we had
  analyzed a different set of characters?
• To answer these questions, we use the statistical
  technique known as bootstrapping

30
Bootstrapping Trees 2

• Suppose the actual data sample consists of n
  observations
• To bootstrap, draw a new sample of n
  observations from the actual data, with
  replacement, and re-analyze the bootstrap sample
• Repeat many times (1,000s) the process of
  drawing a bootstrap sample and analyzing it
• For a phenogram or cladogram, the data that are
  bootstrapped are the characters in other words,
  we re-sample the characters that we analyze to
  make the tree
• Tests with known phylogenies (lab experiments)
  indicate that bootstrap support of 70 or better
  is usually associated with the true phylogeny.

31
Co-speciation of aphids and bacterial symbionts
(Fig. 14.14)
32
Using phylogenies to test evolutionary hypotheses

• Co-speciation do parasites speciate when their
  hosts speciate (vertical speciation), or do
  parasites speciate by lateral transfer to a new
  host (horizontal speciation)?
• What is the order of evolution of adaptations?
• Does continental drift explain the pattern of
  speciation in a taxon?

33

• Families that include eusocial species are
  indicated in boldface type

Sociality and nesting behavior in hymenoptera
(Hunt 1999) (Fig. 11.13
34
Phylogeography of Chameleons (Fig. 14.13)

• Separation of Gondwanaland
• Upper graph is phylogeny of chameleons based on
  sequence of separation of southern continents
  (vicariance hypothesis)
• Lower graph is phylogeny estimated from
  morphological, behavioral, and molecular data
  (Raxworthy et al. 2002). This tree implies that
  chameleons have dispersed from Madagascar to
  Africa on several occasions, from Madagascar to
  the Seychelles, and from Africa to India
• The dispersal hypothesis is supported by the
  presence of chameleons on Reunion and the Comoros
  Is., which are volcanic and have never been in
  contact with continental land masses.

35
Phylo-geography of Chame-leons(Fig. 14.13c)
36
Are ungulates mono-phyletic? (Fig. 14.16)

• According to this figure, are the ungulates
  (artiodactyls and perissodactyls) a monophyletic
  group?

37
Are ungulates mono-phyletic? (Fig. 14.16)

• According to this figure, the ungulates
  (artiodactyls and perissodactyls) are a
  paraphyletic group
• Hooves gained
• Hooves lost

38
Are ungulates mono-phyletic? (Fig. 14.16)

• According to this figure, the ungulates
  (artiodactyls and perissodactyls) are a
  polyphyletic group
• Hooves gained

39
Are these trees different? (Fig. 14.17)