Comparative methods Flashcards
Describe Intra-specific studies
- Single species = restricted variation
- Experiments can create extra variation (e.g change environment)
- However, often not possible to test evolutionary/ecological theories on a single species
Describe why the comparative method involves inter-specific comparisons
- Group of species contain more variation than single species
- Also experienced a bigger range of selective pressures
- Use data collected across lots of species = generate variation
Describe Tinbergen et al. 1962/3 studies into Black headed gulls and Kittiwake
- Black-headed gulls face high predation risk due to ground nesting, leading to adaptive behaviors like eggshell removal and mobbing.
- Kittiwakes, with lower predation pressure, lack these behaviors
- By comparing related species with different ecological contexts, it is a classic example of the comparative method
What is correlated evolution?
How is it analyzed in the comparative method?
Give an example
Correlated evolution examines whether changes in one trait are associated with changes in another.
- Statistical analyses e.g regression or correlation, are used to determine if observed patterns are likely due to chance.
- E.g investigating the relationship between body size and basal metabolic rate or between brain size and survival in birds
What are some advantages of the comparative method?
- Very general
- Incorporate large amounts of variation (trait and ecological)
- Can utilise existing databases
- Can guide experimental and observational work
What is the “Patterns vs Process” problem in the comparative approach?
- A conceptual issue arising from the observational approach, where patterns seen in data may not accurately reflect the actual evolutionary process.
- Patterns might have evolved through different pathways, requiring knowledge of ecological information about ancestors (e.g., their habitat) rather than relying solely on the current distribution of traits.
Why is statistical independence important in the comparative method, and what issues arise with non-independent data?
Statistical independence important bc most statistical methods assume that samples are independent (one value cannot predict another)
Non-independence (e.g data from the same location, animal, or time) can lead to incorrect degrees of freedom and confounding effects.
These issues can result in misleading conclusions due to hidden variable effects.
Why is phylogeny a problem in the comparative approach?
Phylogenetic dependence means closely related species may share traits due to common ancestry, not adaptation. This can:
- Mask true patterns: Traits might appear shared because of evolutionary history, not environmental adaptation.
- Overestimate or underestimate correlations: Ignoring evolutionary relationships can distort patterns.
- Miss independent evolution: Species might evolve similar traits independently, which could be overlooked without considering phylogeny.
What factors make phylogeny important to control in comparative studies?
1) Phylogenetic niche conservatism
2) Evolutionary lags
3) Different adaptive responses
What is Phylogenetic niche conservatism? Give an example
Tendency of species to retain ancestral ecological traits
e.g Darwins Finches
e.g New World monkeys, such as Cebus (capuchins) and Ateles (spider monkeys)
What are evolutionary time lags?
Give an example
Environmental changes occur faster than evolutionary adaptations
E.g Vestigial organs (present but not functional)
E.g Sea temps rising at a rate faster than coral reefs can adapt
What are Different adaptive responses?
Give an example
Species may evolve different solutions to similar challenges.
E.g E.g Macaw (frugivore) vs Kea (scavenger): Same beak morphology
What are the assumptions of the Brownian model
- Traits change all the time i.e evolution always happening
- On av evolution is neutral: traits inc and dec but decreases are as likely as increases
- Rate of evolution is constant: rate of accumulation of diffs from ancestors is always the same
- Species are independent of each other
What are the limitations of the Brownian model?
- Variance increases without bound
- Rate is constant
- Species are independent once they have split
- No ecological mechanisms
What is the Brownian model of evolution?
- The Brownian model of evolution describes trait evolution as a random walk process over time.
- It assumes that trait changes occur gradually and randomly, with no directional bias, leading to increasing divergence among species proportional to their time since divergence.
How is the maximum likelihood method applied to comparative analysis?
- Requires trait data and a phylogeny.
- Different values of λ (phylogenetic signal) are tested to see which best fits the observed data.
- The value of λ that maximizes the likelihood of the data under the model is selected as the best fit.
What is phylogenetic signal?
- Measures the tendency of related species to resemble each other more than species drawn at random
- Quantifies how much a trait’s distribution across species can be explained by their shared evolutionary history.
What does strong phylogenetic signal show?
Indicates that closely related species have similar traits due to shared ancestry
What does a larger maximum liklihood value indicate?
A better fit to the data
(0= no fit, 1 = best fit)
What is λ in the Brownian model, and how is it used?
- λ = parameter that measures the strength of phylogenetic dependence.
- Useful for testing phylogenetic dependence in traits, providing evidence of niche conservatism.
- By allowing for varying levels of phylogenetic dependence in statistical models, λ helps refine analyses to better fit the data.
What does the Ornstein-Uhlenbeck model assume about trait evolution?
- Assumes traits evolve towards an optimal value, with deviations from this optimum being pushed back to the equilibrium value.
- Suggests traits stabilize around an optimum (rather than changing randomly like the Brownian model)
What does the parameter α represent in the Ornstein-Uhlenbeck model?
- α measures the strength of evolutionary constraint.
- If α = 0, there’s no constraint, and trait evolution follows a Brownian motion pattern.
- If α > 0, theres a constraint, with traits being pulled toward a specific optimum value, showing evidence of evolutionary stabilization
Give an example of applying the OU model
(Hansen 1997)
- tells us the evo of body size in horses is non-Brownian
- strong adaptive constraint
- small/large body sizes non-optimal
dont need to remember the ref
Why is spatial location important in the comparative method?
- Spatial location influences the environment, which can drive differences in trait values.
- Species that live closer together are often more similar due to shared environmental conditions and phylogenetic relatedness.
Is the comparative method considered experimental or observational? explain why
- Experimental → ‘gold standard’, manipulate the factor of integers, directly test hypotheses
- Observational → ‘cheap’, use existing and easily collected data. Data is taken directly from nature, so is representative
Comparative is an observational approach - there are drawbacks and typically it is not thought of as a predictive method
But we can use comparative methods fit models to data ad we can use these models to make predictions
How can phylogenetic relationships help make predictions?
- Phylogenetically close species are often similar due to shared ancestry.
- By knowing trait values for one species, we can predict traits for related species.
- This is especially useful when dealing with missing data in large databases
How can comparative methods assist in estimating ecological demographic rates?
- Demographic rates (e.g., population growth) are difficult to estimate due to high data demands.
- James et al. (2020) used phylogenetic imputation to predict demographic rates in birds.
- Predictions were tested by comparing with known data.
- Phylogeny improved predictions compared to simple statistical methods.
- Useful for conservation decisions.
How do geography and phylogeny help predict traits?
- Spatial location influences the environment, which impacts traits.
- Environmental distances often scale with spatial distances.
- Both phylogeny and spatial information can be used to predict traits.
- Spatial data enhances predictive accuracy.
How is the comparative method applied to the IUCN status?
- IUCN 5 point scale of threat
- But, ‘data-deficient’ species (>10% of mammals) are not assessed
- Comparative analysis fills these gaps and shows data deficient species more threatened than assessed species
- Unsuprising bc factors that make species ‘data deficient’ also make them more likely to be threatened
What are the reasons for phylogenetic dependence?
- Phylogenetic niche conservatism
- Evolutionary lags
- Different adaptive responses
What is phylogenetic dependence and why does it matter?
- Closely related species share traits due to common ancestry, making them statistically dependent.
- Shared Ancestry: Traits often inherited from a common ancestor.
- Non-Independence: Observations across species are not fully independent, affecting statistical analyses.
- Phylogenetic Signal: Measures how much trait variation is explained by phylogenetic relationships.
Would you find stronger evidence for phylogenetic signal in ecological or behavioural datasets?
- Higher incidence for ecological (conservative traits)
- Lower for behavioural (labile traits) = Behaviour evolves so organisms can ‘escape’ their evo history
How do we deal with phylogenetic dependence?
- Mathematical models of trait evolution
- Predict how similar/dissimilar sp are as a function of the evo distance between them
- Can then perform stat analysis, knowing how much similarity we expect
- Most commonly used model is Brownian
The Brownian model predicts that the amount of evolution is _________ to the _____ of the branches of the _____
Proportional
Length
Phylogeny
Give some examples of comparative methods
- Phylogenetic contrasts
- Phylogenetic regressions
- Geographical models
- Lamda method
- Ecological interactions
