-
What are assumptions of Tajima's D?
D=X/(√var(x))
X = k – (s / (1 + 1/2 + 1/3 + ... + 1/n-1))
- Constant population size (i.e. equilibrium)
- Infinite-sites compatible, no recombination, no migration
-
What are interpretation of Tajima's D?
D=X/(√var(x))
X = k – (s / (1 + 1/2 + 1/3 + ... + 1/n-1))
- If D is significantly different than zero reject hypothesis of neutrality.
- General rule of thumb - values greater than +2 or less than -2 are likely to be significant.
-
What are possible explanations of Tajima's D?
D=X/(√var(x))
X = k – (s / (1 + 1/2 + 1/3 + ... + 1/n-1))
- Selection acting on beneficial mutations (D > 0) with an excess of intermediate-frequency alleles increasing genetic variation (e.g., balancing selection) that benefit the population, sudden population contraction, or population subdivision.
- Selection removing or maintaining deleterious (i.e. lower fitness) mutations (D < 0) at a low frequency (e.g., population expansion, background selection, purifying or negative selection or selection sweeps)
-
θ estimator (Inheritance scalar):
autosomal/diploid
haploid
X-linked
Y-linked
- Inheritance scalar
- autosomal = 1
- haploid = 0.50
- X-linked = 0.75
- Y-linked = 0.25
-
θ estimator (equations):
autosomal/diploid
haploid
X-linked
Y-linked
- Equations
- autosomal/diploid... θ = 4Ne µ
- haploid... θ = 2Ne µ (haploid)
- X-linked... θ = 3Ne µ (X-linked)
- Y-linked... θ = Ne µ (mt or Y locus)
-
What is a possible explanation of D >0?
Selection acting on beneficial mutations (D > 0) with an excess of intermediate-frequency alleles increasing genetic variation (e.g., balancing selection) that benefit the population, sudden population contraction, or population subdivision.
-
What is a possible explanation of D < 0?
Selection removing or maintaining deleterious (i.e. lower fitness) mutations (D < 0) at a low frequency (e.g., population expansion, background selection, purifying or negative selection or selection sweeps).
-
How can you interpret Tajima's D?
If D is significantly different than zero reject hypothesis of neutrality. General rule of thumb: values greater than +2 or less than -2 are likely to be significant.
-
When is Fu and Li's D more powerful than Tajima's D?
detecting background selection (e.g., deleterious mutations and purifying or negative selection)
-
When is Fu and Li's Fs more powerful than Tajima's Fs?
detecting background selection (e.g., deleterious mutations and purifying or negative selection)
-
When is Fu and Li's Fs less powerful than Tajima's Fs?
for detecting population growth and genetic hitchhiking (e.g., beneficial mutations and positive selection)
-
When is Fu and Li's D less powerful than Tajima's D?
for detecting population growth and genetic hitchhiking (e.g., beneficial mutations and positive selection)
-
Many software will output both Tajima's and Fu/Li's values of D and F. What is indicated if only Fu and Li's tests are significant?
If only Fu and Li’s tests are significant - suggests background selection
-
Many software will output both Tajima's and Fu/Li's values of D and F. What is indicated if only Tajima's tests are significant?
If only Tajima's D and Fs are significant - suggests population growth or genetic hitchhiking
-
Haplotype networks
- A phylogeographic approach
- statistical parsimony and SplitsTree graphs
- use if you want to assess geographical spread of molecular data without modeling
-
ABC
- A phylogeographic approach (Approximate Bayesian Computation)
- use if you want to model rates of spread and take into account geographical features
-
Coalescent
- A phylogeographic approach (mutation ages)
- use if you want information about population size, migration and geographical origins of mutations
-
Assumptions of Tajima's D
- constant population size (equilibrium)
- Infinite-sites compatible
- no recombination
- no migration
-
Four-gamete test
- method for detecting historical recombination events
- Infers a recombination event under the assumption of infinite-series
- H4 is most likely ancestral because it:
- Shares fewer derived mutations compared to H2 and H3.
- Does not participate in recombination events in the ARG.
- Is in a "simpler" form compared to the recombinant haplotypes
-
infinite-series assumption
- repeat mutations have 0 probability
- parallel evolution unlikely
-
IMgc
- IMgc uses the four-gametes criteria to remove either sequences or variable sites containing evidence of recombination
- It tests for recombination and finds the largest non-recombining partition
-
Ancestral Recombination Graph (ARG)
-
What do different genetic markers provide, in terms of resolution?
population processes on different evolutionary time scales
-
-
How can different genetic markers help to provide regulation of population processes on different evolutionary time scales
The integration of multiple loci comprising mixed markers is more powerful for evolutionary inference than multiple loci of the same genetic marker (e.g., IMa3 allows analysis across multiple data types).
-
How can you reduce stochasticity associated with a single locus?
integrating multiple loci
-
haplotype
a group of genes within an organism that was inherited together from a single parent
-
Exclude recombination haplotypes
- IMrg can do this
- Example with Ancestral recombination graph
-
- The recombination event between sites 2 and 3 is shown in the oval
- H2 gets a prefix from H3 and the suffix from H1
- H4 is the ancestral haplotype
-
What provides resolution of population processes on different evolutionary time scales?
Different genetic markers (SNPs, SSRs, etc)
-
How can you reduce stochasticity associated with single locus inferences?
integrate over multiple loci
|
|