Tree Building

1
Tree Building

• What is a tree ?
• Cladograms
• Trees
• Scenario
• How to build a tree ?
• Observations
• First Principles
• Assumptions
• Methods

2
What is a tree ?

• Cladograms and Trees
• Both are graphs in mathematical terms
• A graph is a collection of nodes (vertices) and
  lines / branches (edges) connecting the nodes.
• A cladogram/tree for our purposes is allowed at
  most one edge between any two vertices.

3
Cladogram Trees - 1

• The degree of a node is the number of branches
  that contain that node.
• A node of degree 1 is called a leaf (or terminal
  node)
• All nodes that are not leaves are called internal.

4
Cladograms Trees - 2

• A tree is elementary if no node has degree 2
  (net-work in cladistic jargon).

• A root is a distinguished node with degree 2,
  locating the start of the tree.

5
Cladograms Trees - 3

• An unrooted tree is binary if every node has
  degree 1 or 3.
• A rooted tree is binary if it has a root of
  degree 2 and every other node has degree 1 or 3.

6
Cladograms Trees - 4
Label
Leaf
Branch
?
Node
Root

• Labeled rooted binary tree

7
Cladograms Trees - 5

• Whats the difference?
• Cladogram
• Cladogenesis
• branching events as indicated by character state
  changes
• Tree
• Anagenesis
• amount and duration of change inference of
  ancestor-descendant relationships

8
A Cladogram is a

• Statement about the distribution of (shared)
  character states.
• Branching diagram depicting nested sets of
  synapomorphies resulting in a summary statement
  of sister-group relations among taxa.

9
Nested sets of Synapomorphies

• Detection of relationships by distribution of
  character-states in species X, Y, and Z.

aBCde
abcde
10
Relationship, and Kind of Groups

• Relationship criterion
• Recency of common ancestry
• A species X is more closely related to another
  species Y than it is to another species Z if, and
  only if, it has at least one stem species in
  common with species Y that is not a stem species
  of Z
• (Hennig, 1966, p.74)
• X and Y are sistergroups.

11
A Phylogenetic Tree is a

• Branching diagram where
• the nodes represent real or hypothetical
  ancestors,
• the branching represents speciation, and
• the branches represent descent with modification.

12
Cladograms Trees - 6

• Cladogram set of trees Every picture tells a
  story

?

13
Cladogram Set of Trees
14
The Cladistic Party Line ...

• There is simply no possible way to distinguish
  ancestors from extinct lineages.
• if something is in fact an ancestor, there are
  no data that can refute the hypothesis that it is
  an extinct lineage and not an ancestor.
• (Mark Siddal, 09/01/96, sci.bio.systematics).

15
and its Counterpart

• A is the ancestor of B is a perfectly valid
  hypothesis, and one that is easily falsified. All
  it would take to falsify it is to find an
  autapomorphy in A that is not found in B.
• (Ron DeBry, 18/01/96, sci.bio.systematics)

16
How to build a cladogram

• Observations
• Character-state distri-butions over taxa (data
  matrix), or derivation thereof (distance matrix)
• First principles
• Assumptions
• Process
• Model
• Data
• Type and Quality

17
First Principles

• Evolution (descent with modification) occurs.
• Evolution results predominantly in a hierarchical
  scheme of relationships among the entities
  involved.
• ...?

18
Assumptions - 1

• The fact that parsimony methods are known to
  fail in reconstructing phylogeny when there are
  unequal rates of evolution, and fail in a
  systematic way (e.g., put long branches together
  when they really should each go with one of the
  short branches) suggest that certain
  conditions of the process of evolution have to be
  met in order for the method to be useful . If
  a method is only useful when certain conditions
  of the evolutionary process are met, I would
  think that these conditions might as well be
  thought of as assumptions.
• (Andrew J. Roger, 08/01/96, sci.bio.systematics)

19
Process

• I have a pretty good idea of how evolution
  works, thus I can check how my data fit these
  ideas.
• Given the phylogeny, what is the probability to
  find the data as I did ?
• Model Statistical Framework
• Maximum likelihood

20
Assumptions - 2

• The philosophical part that deserves more
  explanation is how you get from whatever general
  principles you invoke (parsimony) to the
  specific numerical method used.
• Compatibility methods represent discarding a
  character because it has some sign of conflict
  with others. If there are two kinds of
  characters, really horribly noisy and pretty
  clean, that is a sensible thing to do. If there
  are instead two kinds, pretty clean and a little
  noisy, it is not. So I do not see how the
  principle of parsimony decides in advance which
  of these situations we are facing.
• (Joe Felsenstein, 14/12/95, sci.bio.systematics)

21
Data

• My data will tell me what the optimal set of
  branching events is and from there I will try to
  grasp what actually could have happened.
• Parsimony
• Group / Component Compatibility
• Character Compatibility

22
Observations

• Molecular data
• DNA sequences
• nuclear, mitochondrial, ribosomal
• DNA-DNA hybridization
• Restriction-site and -fragment
• Allelic isozymes
• Morphological data
• Anatomical data
• Chemical data

23
Black Boxes ?
Phylogenetic Trees
Principles - 1
Observations
Assumptions
Assumptions
Optimality
Methods
Cladogram(s)