Transcript 3000_2013_1e

GENE 3000
Fall 2013
slides 127-159
wiki. wiki. wiki.
4 species can be related to each other in only 3 ways
4 species can be related to each other in only 3 ways
B
A
C
evolution - story of
gain
and
loss
• when and how were complex eyes evolved?
in what species are they lost? are the genes
required to develop eyes still there? can
they be expressed in different ways?
•
•
how frequently are genes gained and lost?
how often is there lateral gene transfer
between hosts and pathogens, or
microbes?
why are there more A/T bases in parts of
the mitochondrion?
Nature, June 2012
all starts with this
• phylogeny: visual
representation of the
terms
evolutionary history of populations,
genes, species
• tips: terminal ends of a phylogeny,
representing the populations, genes,
species
• branches: lineages evolving through
time
• node: where lineages split; the common
ancestor to descendants
• clade: a group all descended from one
• phylogeny: visual
representation of the
terms
evolutionary history of populations,
genes, species
• tips: terminal ends of a phylogeny,
representing the populations, genes,
species
• branches: lineages evolving through
time
• node: where lineages split; the common
ancestor to descendants
discovermagazine.com
• clade: a group all descended from one
organisms are related
through descent from
ancestors
progression of generations
more than 2 million
identified and classified
species extant
taxonomy
(systematics)
• groups defined on
previous slides are
clades
• group comprised of
one organism and all
its descendants (monophyly)
• indication of homology
Archosaurs
how do we know?
• inference, reconstruction
• characters are heritable traits that can
be compared (an A in position 117;
body hair)
• careful evaluation of homology
• assume that having more things in
common indicates more recent
common ancestor (longer shared
history)
homology and
homoplasy
• similarity without shared descent is
homoplasy - convergent evolution
(wings) or evolutionary reversal (snakes
don’t have legs)
• homology may require full evidence,
e.g. development of insect wings very
different from development of bird
wings
phylogeny inference
•
•
morphological
characters
(presence, absence,
quality)
DNA-based
characters work the
same (idea is there
may be more
independence, and
more of them)
our data matrix
Text? in NEXUS format
parsimony analysis
• assume each state evolved only once
(not necessarily true)
• principle of parsimony: the simplest
explanation is the most likely
• we can identify how often this
assumption is violated
one character
• tentacular forelimbs might be able to
split into 2 clades, but cannot fully
resolve tree
parsimony is a
CRITERION
• our assumption of parsimony does not
find the tree, it is a score given to a tree
• trees are searched algorithmically for all
possible topologies (or all possible
better than score X)
• a parsimony phylogeny analysis returns
the set of most-parsimonious trees
our tree
eyeless
“presence of 2 eyes” - obviously
negatively correlated with “eyeless”
tail
digits
dark stripes
hmmm correlated with digits!?
wings
circular tail fin
“forked” tail fin
“protruding body”
2 digits - an autapomorphy
our tree
• parsimony tree(s) from class data
matrix
• which characters exhibit HOMOPLASY
• remember: homoplasy may be
evolutionary convergence/reversal, may
also reflect our own uncertainty about
the character and how it develops!
• see Box 1 (chapter 4) for another
example
molecular data
•
we will come to this later how it is analyzed, but
we see the problems of
•
•
•
•
highly correlated, non-independent characters
hard to identify very many characters
thus not a fully resolved phylogeny (but good!)
if you sequence a single gene region, might
have 100 characters that are unbiased by the
person collecting the data
example of
carnivores
define “carnivore” please...
why study phylogeny
of carnivores?
lets ask a simple question:
did aquatic pinnipeds evolve
only once, or were there
multiple transitions to the
sea?
12 characters, 10
carnivores
synapomorphies
13
13
20 characters, 3 equally
parsimonious trees
our result is robust to
this uncertainty: one origin of
pinnipeds (consensus at left)
parsimony
• simple assumption, simple model - may
lead to several ‘equally parsimonious’
results
• more data will not always solve the
problem
• more complex models evaluate the
non-independence of data, the
empirical patterns of DNA substitution,
and probability theory