Recombination

1
Recombination
2
What is Recombination?
Crossing over
Recombinants
Parental Sequences
Parental sequences and recombinants
Obviously both nucleic acid strands must be in
the same cell
3
How Does Recombination Occur?
Strand invasion
Broken genome
Recombinants
Repair
Mechanisms Double stranded break and repair - eg
DNA Viruses
4
How Does Recombination Occur?
Disintegrated genomes
Recombinants
Reassembly
Mechanisms Double stranded break and repair - eg
DNA Viruses Disintegration and repair - eg human
influenza virus
5
How Does Recombination Occur?
Template switch during RT
Recombinant
Mechanisms Double stranded break and repair - eg
DNA Viruses Disintegration and repair - eg human
influenza virus Template switching during reverse
transcription - eg HIV
6
Why is Recombination Important?
Almost all genomes undergo recombination Why do
they bother?
An artifact of double stranded break repair?
A very effective way of concentrating beneficial
mutations in single genomes?
Recombination gives every allele the opportunity
to determine its own destiny. Recombination
allows higher mutation rates more aggressive
exploration of sequence space. Recombination
enables fitness peak jumpingand provides an
access route to island peaks.
Fold increase in activity
Mutation alone
Mutation and shuffling
7
Why is Recombination Important?
It is important to population geneticists
Recombination moves chunks of sequence between
genomes. Within chunks alleles with high fitness
value will become fixed along with neutral
alleles that hitchhiked a ride on the same
chunk. Indirect selection of neutral alleles
results in decreased variability of neutral
alleles. Decreased variability (1) decreases
estimates of effective population sizes (2) could
be interpreted by population genetic tests of
natural selection as evidence of either recent
selective sweeps or recent population expansion.
8
Why is Recombination Important?
It is important to phylogeneticists
Recombination allows genomic regions to have
different evolutionary histories - ie no single
phylogenetic tree can describe the ancestry of
recombining sequences. This complicates/prevents
effective use of phylogenies in tracing routes
of disease transmission, determining molecular
clock rates, estimating mutation bias and rate
heterogeneity, and identifying sites under
positive selection. Recombination may
compromise guide tree based alignment methods.
9
Effect of Recombination on Branch Lengths
tMRCA Region 2
tMRCA Combined
tMRCA Region 1
A
B
Tree for region 1
C
D
Breakpoint
A
Tree for region 1
B
C
D
A
B
Tree for combined regions
C
D
From Awadalla, 2003
10
Effect of recombination On Interpretation of Tree
Shapes
Terminal branches the same length as internal
branches. This tree would be expected if most
mutations were neutral and the population size
was constant. Eg Papilloma viruses
Terminal branches longer than internal branches.
This tree would be expected if either the
population size was expanding or recombination
was rampant. Eg HIV and Foot and mouth disease
virus
11
Effect of Recombination on Tests for Mutation
AAAAAAAAAAAAAAA
P1
AAAAAAAAAACCCCC
Re
AAAAAAAAAAAAAAA
X
CCCCCCCCCCCCCCC
P2
Homoplasies will be detected at these sites ie
these sites will look like they have been subject
to recurrent mutations They will appear to be
more mutable than they really are. Mutation-rate
heterogeneity over the entire sequence will be
overestimated. Molecular clock hypothesis might
be falsely rejected. Timing of events
underestimated due to overestimation of terminal
branch lengths.
Effect of recombination on tests for selection?
12
Recombination Detection Methods
Non - Parametric (Dont use population genetic
models)
Parametric (Model based)
Look at linkages between alleles in a
population.
Look at phylogenetically informative signals
Identify breakpoint positions, recombinants and
parentals
Estimate population recombination rates
Bootscanning, Sister Scanning, Maximum Chi
Square, GENECONV, RDP, Reticulate, and a lot of
others.
LDHat, Sites
13
Method Performance
Power
Homoplasy
MaxChi
Geneconv
RDP
Simulated
Recpars
Real
Bootscan
False Positives
PIST
After Posada and Crendall 2001 and Posada 2002
Can methods detect recombination?
14
RDP Version 2.0
Applies 8 non-parametric methods including, 5
methods for automated recombinant detection RDP,
GENECONV, Maximum Chi Square (MaxChi),Bootscan
Sister Scan (SiScan)
Triplet (T) or Doublet (D)
Variable (V) or All (A) sites
Method
Sliding Window
RDP
T

V
-
GENECONV
D
V
MaxChi
T/D
V


Bootscan
T
A

SiScan
T
A
15
Triplet scanning methods
P1
R
P2
VS
RDP
Identity
VS
Chi Value
MaxChi
BS Support
AS
Bootscan
SiScan
AS
Z-Value
R - P1
R - P2
P1 - P2
16
Conclusions
Recombination can seriously compromise nearly
every type of analysis that requires aligned
sequences. Many different methods of
recombination detection are available. Only a
few of these estimate population recombination
rates, identify recombinants/parentals, or
identify breakpoint positions. The power of
different methods to detect the presence of
recombination varies. You should always choose a
powerful method that does not detect too many
false positives - It is unknown how methods
compare WRT breakpoint or daughter
identification. Combinations of methods are
preferable.
17
Recombination Detection
18
Recombination Detection
19
Recombination Detection
20
Recombination Detection
21
Recombination Detection
22
Recombination Detection
23
Recombination Detection