Biology Reference
In-Depth Information
(b)
25
(a)
C
3
20
(x, y)
D
(20, 120)
15
B
C
1
(18, 115)
E
(16, 110)
C
1
10
A
C
3
F
(4, 87)
C
2
5
C
2
C
(5, 90)
G
(6, 93)
2
4
6
8
10
12
14
Contrasts in x
Fig. 2.8
(a) A simple phylogeny. In this tree, A gave rise to two descendants, B and C,
each of which gave rise to two more descendants: D and E, and F and G. The character
states (x, y) of the four living species (D, E, F, G) are measured. Those of their ancestors
(B and C) have to be estimated. In this case they are assumed to have values
intermediate between those of their descendants. There are three independent
contrasts in this tree: C
1
, C
2
and C
3
. (b) Plotting contrasts in x against contrasts in y
shows that there has been correlated evolution in these two traits.
The character states (diet, brain size, mating system and so on) of extant species can
be measured. How can we know what their extinct ancestors were like? For morphological
traits, fossils can be useful but behavioural traits rarely leave a fossil record. We have,
therefore, to make an educated guess about ancestral states. Again, various statistical
methods are available. The simplest method is parsimony: assign ancestral states to
minimize the number of evolutionary changes in the tree from ancestral to extant
species. More complex methods are maximum likelihood and Bayesian statistics, which
consider which are the most likely ancestral states among various possibilities. In
general, if there are frequent changes in the tree then ancestor reconstructions become
more uncertain, especially for more distant ancestors (Schluter
et al
., 1997).
Estimating
ancestral states
Independent contrasts
Joe Felsenstein (1985) introduced the method of independent contrasts to solve the
problem that species are not independent. Fig. 2.8 is a simple example to explain the
method. The key point is that we can assume that two species have evolved independently
since their divergence. Therefore, their degree of divergence is independent (statistically)
from other changes in the tree. These divergences between related taxa provide independent
changes, or contrasts, for our analysis (D versus E and F versus G in Fig. 2.8a). In addition
to comparing pairs of species at the tips of the tree, we can also work backwards and
compare at higher levels, effectively comparing groups of species or ancestors (B versus C).
This is often done by assuming that ancestral values of continuous traits (e.g. brain size)
Using phylogenies
to identify
independent
evolutionary
changes …
