Download my_phylogeny1

Phylogenetic analysis
taken from
http://allserv.rug.ac.be/~avierstr
and
http://www.cs.otago.ac.nz/cosc348/Lectur
es/MSAPhylogeny.htm
And
Introduction to Bioinformatics course
slides
Purpose of phylogenetics :
• Reconstruct the evolutionary relationship between species
Experience learns that closely related organisms have similar
sequences, more distantly related organisms have more dissimilar
sequences.
• Estimate the time of divergence between two organisms since they
last shared a common ancestor.
But…
• The theory and practical applications of the different models are not
universally accepted.
• Important to have a good alignment to start with. (Garbage in,
Garbage out)
• Trees based on an alignment of a gene represent the relationship
between genes and this is not necessarily the same relationship as
between the whole organisms. If trees are calculated based on
different genes from organisms, it is possible that these trees result in
different relationships.
Why is phylogeny imporant
• Determining tree of life (e.g., for a new
organism)
• Determining gene function
• Understand which parts of the
gene/regulatory sequences are important
• Tracing the evolution of genes – horizontal
gene transfer etc.
Protein or DNA?
• As with Multiple Sequence Alignment –
proteins are preferred
– More informative
– Shorter in length
– Less chance of multiple mutations at the
same site
• When DNA?
– A non-coding sequence
– Proteins too similar
Terminology :
•
•
•
•
•
•
node : a node represents a
taxonomic unit. This can be a
taxon (an existing species) or
an ancestor (unknown
species : represents the
ancestor of 2 or more
species).
branch : defines the
relationship between the taxa
in terms of descent and
ancestry.
topology : is the branching
pattern.
branch length : often
represents the number of
changes that have occurred
in that branch.
root : is the common
ancestor of all taxa.
distance scale : scale which
represents the number of
differences between
sequences (e.g. 0.1 means
10 % diff
Possible ways of drawing a tree :
Unscaled branches : the length is not proportional to the number
of changes.
Possible ways of drawing a tree :
•Scaled branches : the length of the branch is proportional to the
number of changes (usually in PAMs). The distance between 2
species is the sum of the length of all branches connecting them.
Possible ways of drawing a tree :
• Rooted trees: the root is the common ancestor. The direction of
each path from the root corresponds to evolutionary time.
• Unrooted tree: specifies the relationships among species and
does not define the evolutionary path.
Rooted vs. unrooted trees
3
1
2
3
1
2
9
Rooted vs. Unrooted.
The position of the root does not affect the
MP score.
10
Intuition why rooting doesn’t change the score
Gene number 1
1 or 0
1
0
1
s1
s4
s3
s2
s5
1
1
1
0
0
The change will always be on the same branch, no matter where the
root is positioned…
11
We want rooted trees!
How can we root the tree?
12
13
14
Gorilla gorilla
(Gorilla)
Pan troglodytes
(Chimpanzee)
Homo sapiens
(human)
Gallus gallus
(chicken)
15
Evaluate all 3 possible UNROOTED trees:
Human
Chicken
Human
Chimp
Gorilla
Human
Chicken
Gorilla
Chimp
Chimp
MP tree
Gorilla
Chicken
16
Rooting based on a priori knowledge:
Human
Chimp
Chicken
Gorilla
Chicken
Gorilla
Human Chimp
17
Ingroup / Outgroup:
Chicken
OUTGROUP
Gorilla
Human Chimp
INGROUP
18
Tree of life
Distance-based methods
• Compress all of the individual differences between pairs
of sequences into a single number – the distance.
• Starting from an alignment, pairwise distances are
calculated between DNA sequences as the sum of all
base pair differences between two sequences (the most
similar sequences are assumed to be closely related.
This creates a distance matrix.
• From the obtained distance matrix, a phylogenetic tree is
calculated with clustering algorithms. These cluster
methods construct a tree by linking the least distant pair
of taxa, followed by successively linking more distant
taxa.
• Algorithms: UPGMA clustering , Neighbor Joining.
• Assumes molecular clock
ClustalW!
Cladistic methods
• Trees are calculated by considering the various possible
pathways of evolution and are based on parsimony or
likelihood methods. These methods use each alignment
position as evolutionary information to build a tree.
• Parsimony : Looks for the most parsimonious tree: the tree
with the fewest evolutionary changes for all sequences to
derive from a common ancestor.
Phylip
• Slower than distance methods.
• Assumes molecular clock
• Maximum Likelihood : Looks for the tree with the maximum
likelihood: the most probable tree.
• this is the slowest method of all but seems to give the best
result and the most information about the tree.
Phylip
• No molecular clock assumption
Even the best
evolutionary
models can't
solve this
problem...
Two homologous DNA sequences which descended from an ancestral sequence and
accumulated mutations since their divergence from each other. Note that although 12
mutations have accumulated, differences can be detected at only three nucleotide sites.
Molecular clocks
Dickerson, 1971
• Assumption:
constant rate of
evolution
• Different rate for
different genes:
Millions of years since divergence
Human insulin
Insulin multiple alignment
Problems with molecular clocks
Surprisingly, insulin from the guinea pig evolved seven
times faster than insulin from other species. Why?
The answer is that guinea pig insulin does not bind two
zinc ions, while insulin molecules from most other species
do. There was a relaxation on the structural constraints of
these molecules, and so the genes diverged rapidly.
Building trees with ClustalW
http://www.ebi.ac.uk/clustalw/
Place
alignment
here
Choose a
tree here
PHYLIP
• A suite of phylogeny tools
• Both web servers and stand-alone
applications
• Used for distance/parsimony/maximum
likelihood
• http://bioweb.pasteur.fr/seqanal/phylogeny
/phylip-uk.html
Sequences
Bootstrapping
• Assigns confidence to individual tree
branches
• Columns of the alignment are randomly
sampled (with replacement) and the tree is
recomputed X many interactions
• Boorstrap value of a branch = how many
iterations had it.
Collections of homologous genes
• Homologene @ Entrez
– http://www.ncbi.nlm.nih.gov/sites/entrez?db=homolog
ene
• COG – Clusters of Orthologous Genes
– Results of Blast All-vs-All between genomes. Genes
within the same COG are “pairwise best hits”
– http://www.ncbi.nlm.nih.gov/COG/
• RDP – Ribosomal sequences
– The “standard” sequences for doing species
phylogeny
– Focused on Bacteria
– http://rdp8.cme.msu.edu/html/
Orthologs
Homologous sequences are orthologous if they
were separated by a speciation event:
If a gene exists in a species, and that species
diverges into two species, then the copies of
this gene in the resulting species are
orthologous.
32
Orthologs
• Orthologs will typically have the same or
similar function in the course of evolution.
• Identification of orthologs is critical for
reliable prediction of gene function in
newly sequenced genomes.
33
Orthologs
ancestor
a
speciation
a
descendant 1 (e.g., human)
a
descendant 2 (e.g., dog)
34
Paralogs
Homologous sequences are paralogous
if they were separated by a gene
duplication event:
If a gene in an organism is duplicated,
then the two copies are paralogous.
35
Paralogs
• Orthologs will typically have the same or
similar function.
• This is not always true for paralogs due to
lack of the original selective pressure upon
one copy of the duplicated gene, this copy
is free to mutate and acquire new
functions.
36
Paralogs
a
Duplication
a
b
37
(taken from NCBI)
38
Using BLAST and phylogeny to
study gene evolution
39
Mol. Biol. Evol. (2005) 22:598-606
40
Evolutionary rate and conservation
Functionally or structurally important sites are
conserved:
Conserved sites
Variable sites
 “slow” evolving sites
 “fast evolving” sites
Sites which are under a functional/structural
constraint are conserved, and evolve slowly
41
Conservation in an MSA
S1
S2
S3
KITAYCELARTDMKLGLDFYKGVSLANWVCLAKWESGYN
MPFERCELARTLKRMADADIRGVSLANWVCLAKWFWDGG
MPFERCELARTLKRMMDADIRGVSLANWVCLAKWFWDGG
From the MSA (and the tree), one can determine how
conserved is a gene.
42
“Inverse relation between
evolutionary rate and age of
mammalian genes”:
Protocol
43
Step 1 - BLAST
Build the dataset of mammalian
genes
44
Step 1 – BLAST: build the dataset
of mammalian genes, based on
mouse-human ortholog pairs
• The orthologs are defined as pairs of
reciprocal BLAST hits.
• Eliminate genes with more than one
potential orthologous sequence.
• Select only genes which the human
protein was functionally annotated.
45
Step 2 – Calculate conservation
46
Step 2 – Calculate Evolutionary
Rates (Conservation)
For each orthologous pair:
• Alignment at the amino acid level.
• Measure evolutionary rate
The dataset contained 6,776 human-mouse
gene pairs.
47
Step 3 – Assignment of Temporal
Categories
How old is each gene? Used BLAST to
find homologs in 6 different eukaryotic
genomes
48
Caenorhabditis
elegans
Drosophila
melanogaster
Takifugu rubripes
Schizosaccharomyces
pombe
Arabidopsis thaliana
Saccharomyces cerevisiae
49
What is Old ?
• Presence of any
homolog in all the
6 genomes.
What is Presence ?
 Using an e-value cutoff of
10-4 in BLAST.
OLD
METAZOANS
DEUTEROSTOMES
TETRAPODS
50
• METAZOANS - Organisms whose bodies consist of
many cells, as distinct from Protozoa, which are
unicellular; also commonly called animals.
• DEUTEROSTOMES - The second of the two main
groups of bilaterally symmetrical animals. The name
derives from 'deutero' (second) 'stome' (mouth), referring
to the origin of the definitive mouth as an opening
independent from the blastopore of the embryo.
• TETRAPODS - Any four-legged animals, including
mammals, birds, reptiles and amphibians.
51
Human
Tetrapods
Mouse
Fish Deuterostomes
Insect
Metazoa
Worm
Yeast
Old
(eukaryotes)
Plant
52
Results
53
Negative correlation between “age” of genes and the rate of evolution
Negative correlation
between “age” of genes
and the rate of evolution
Evolutionary rate
Evolutionary rate
54
Evolutionary rate
Evolutionary rate
Control.
• Changing the sensitivity of the BLAST
detection to a more conservative one of 10-10,
did not significantly affect the result.
55
Explanations
56
• Functional constraints remained constant throughout the
evolutionary history of each gene, but the newer genes
are less constrained than older genes.
• Functional constraints are not constant, rather they are
weak at the time of origin of a gene and they become
progressively more stringent with age.
57
Eran Elhaik, Niv Sabath,
and Dan Graur
58
Mol. Biol. Evol. 23(1):1–3. 2006
Goal
• To show that these results are an artifact
caused by our inability to detect similarity
when genetic distances are large.
59
Simulation
60
The evolutionary process
Ala
Arg
…
Va
l
Ala
Arg
…
Replacement
probabilities
Val
Rat
Mouse
Cat
Dog
61
Fly
The evolutionary process
Ala
Arg
Va
l
…
Ala
Arg
…
Replacement
probabilities
Val
Rat
Mouse
Cat
V
Dog
62
Fly
The evolutionary process
Ala
Arg
Va
l
…
Ala
Arg
…
Replacement
probabilities
Val
Rat
V
Mouse
Cat
V
Dog
63
Fly
The evolutionary process
Ala
Arg
Va
l
…
Ala
Arg
…
Replacement
probabilities
Val
Rat
L
V
Mouse
Cat
V
Dog
64
Fly
The evolutionary process
Ala
Arg
Va
l
…
Ala
Arg
…
Replacement
probabilities
Val
L
L
L
Rat
L
Mouse
I
Cat
M
Dog
V
V
V
65
Fly
The evolutionary process
And repeat the process for all positions…
(assume: each position evolves independently)
Rat
Mouse
Cat
Dog
Fly
L
L
I
M
V
M
M
M
M
M
T
T
T
T
T
G
G
G
G
G
S
S
S
S
S
H
G
H
G
W
M
M
I
I
R
G
A
G
G
G
N
N
Y
L
R
F
H
A
T
M
I
V
M
R
Y
I
I
F ...
A
A
66
The aim of the simulations: generate sequences with the following phylogenetic
relationships:
All the genes
originated in the
common ancestor
of A,B,C,D,E and
are, thus, of equal
age.
Remote homologs
from increasingly
distant taxa
(similar to fish,
insect, yeast…)
Similar to the
human and
mouse
orthologous
genes.
A
B
C
D
E
67
Simulation
• They simulated genes with 101 different
rates.
• High rate  higher likelihood for an amino
acid replacement in each branch.
68
After simulating the sequences:
Use BLAST, at the same way that Alba and
Castresana used it, to detect homology
between gene A to genes C,D and E.
69
Only one difference –
the groups names
OLD
SENIORS
METAZOANS
ADULTS
DEUTEROSTOMES
TEENAGERS
TETRAPODS
TODDLERS
70
Results
71
Same as Alba and Castresana
72
But all the simulated
genes are at the same
“age”.
What is the problem ???
73
We can only count genes
that are identified as
homologous by the
protocol … BLAST
74
Alba and Castresana
may have, thus, failed to
spot the vast majority of
homologs from among
the fastest evolving
genes
75
The vast majority of the fastest evolving genes are undetectable
even when the cutoffs are extremely permissive.
76
Conclusion
77
The inverse relationship between
evolutionary rate and gene age is an
artifact caused by our inability to detect
similarity when genetic distances are
large.
78
• Since genetic distance increases with time
of divergence and rate of evolution, it is
difficult to identify homologs of fast
evolving genes in distantly related taxa.
• Thus, fast evolving genes may be
misclassified as “new”.
79
So, the only conclusion
Slowly
evolving
that
can be
drawn from
genes
Alba
and Castresana’s
study
is that
evolve
slowly
!!!
80