Macroevolution

Macroevolution_Lecture_2026.pdf

Intro

Macroevolution studies evolution above the species level, focusing on the tree of life in a monophyletic clade

Mostly concerned with Phylogenetics and Speciation
How has lineage diversification occured?
How have traits evolved along clade history?
Where (geographic) were the origins of the lineage?

Example - global phylogeny of butterflies

(Kawahara et al 2023)
Looks at host species, diversification rate across a giant tree of all butterflies and made a really cool chart:
!Macroevolution_Lecture_2026, p.4

Phylogenetic signal

Trait values are not independent because of species relatedness

Phylogenetic signal is the tendency of related species to resemble each other more than a species drawn at random from the same tree

Assumes that traits evolve following Brownian Motion:

Models of trait evolution

Brownian Motion

This is a model for particle movements in physics that is also used as the basic model for trait evolution. The trait value will change according to a defined rate based on a Normal Distribution with $μ = 0$ and defined variance.

d X (t) = σ d W (t)

It assumes that the rate of change is constant through time. Therefore, trait value difference between two species is dependent on time since speciation (divergence) and the rate of change/variation, according to the formula above
!Macroevolution_Lecture_2026, p.13

Assuming Brownian motion, we expect that closely related species will be more phenotypically similar than distantly related species (makes sense)

This does not require the presence of selection or adaptation driving the variation in traits among specis
Adaptation (Natural Selection) may obscure the phylogenetic signal

Tests of phylogenetic signal

Pagel's $λ$
- Varies from 0 to 1, $λ = 1$ trait evolution is what we would expect from pure Brownian motion, lower values indicate selection is taking place
- maximum likelihood estimate is usually used to test this
Blomberg's K:
- $K < 1$ - less similar than expected under BM,
- $K = 1$ is pure BM
- $K > 1$ more similar than expected under BM
- Usually tested with permutation test
Moran's I, used in Spatial autocorrelation, is also used to detect autocorrelation in phylogenetic signal
These tests are sensitive to species number in the phylogeny.

Ornstein-Uhlenbeck (OU) model of evolution

Evolution with an attraction to an optimal value
!Macroevolution_Lecture_2026, p.22
Selection strength is proportional to distance from the optimum value
If there is no selection strength, variation is only due to Brownian motion (random)
Higher selection strength ( $α$ ) will change the direction of the divergence, so the final values won't be distributed around the mean from Brownian motion:
!Macroevolution_Lecture_2026, p.30
If selection strength is high enough, then this can erase/mask the phylogenetic signal, leading to convergent evolution or Cryptic Species

Summary

Brownian motion (BM) is a random deviation of trait values that depend on time since speciation and rate of variation
OU model is a variation of BM model that includes selection strength and optimum phenotype/niche
Phylogenetic signal assumes brownian motion, which can be masked by high enough selection

Diversification Rates

Group discussion

From the clade point of view, diversification rates are higher when:

Groups that have experienced adaptive radiation events
Generation times are shorter
Mutation rates are higher
Species interaction or specialization - trophic interactions, insects that specialize on certain plant species or parasitism
From a regional point of view:
Less Population Genetics - Migration
Latitudinal Diversity Gradient
- Solar radiation, temperature,
Island Biogeography - size of the region/island
More niches available in region

Processes affecting diversity:

Speciation, extinction and dispersal

Speciation

Speciation#Allopatric speciation
- Reproductive isolation
Speciation#Sympatric speciation
- Ecological speciation
See Causes of speciation
Adaptive Radiation: The idea of a new phenotype that can suddenly fill/explore an empty niche
- an "early burst" model of evolution - there is a sudden peak in rate of evolution that slows down as the niche is filled

Rates of speciation and extinction can be:

Clade dependent (different clades have different history)
Time dependent (diversification rate changes over time)
State-dependent (diversification rate changes with phenotype evolution or in a different biogeographic region)

Birth-death model

Net diversification rate: R = speciation per lineage - extinction per lineage: $R = S - E$
Relative extinction rate: Z = extinction per lineage / speciation per lineage: $Z = E / S$

In terms of number of species $N$ at time $t$ :

N_{t} = N_{0} \cdot e^{t (S - E)}

Therefore we can estimate $R = S - E$ with:

R = \frac{\log (N_{t}) - \log (N_{0})}{t}

Applying this to a phylogenetic tree, where 8 species diverged from 2 over 5 My:
!Macroevolution_Lecture_2026, p.44

R = \frac{\log (8) - \log (2)}{5} = 0.2772 species/My

Since we are not considering extinct species, this is a "pure birth" model.

Tip-based rates

Lots of different metics, focusing on length of lines:
- Freckleton et al (2008) - number of nodes per time (Node density)
- Jetz et al (2012)- inverse of equal splits (ES) - $D R = 1 / E S$
- !Macroevolution_Lecture_2026, p.50
- Bayesian estimation of speciation and extinction (BAMM): Rabosky (2014), this allows us to estimate extinction rates.
Advantage - by focusing only on tips, we can spatialize the results and plot them into maps, since we know the distribution of the extant species

State-dependent Speciation Extintion models

This class of models incorporates the influence of traits in diversification rates
This one and the BAMM (Bayesian) model are the only ones allow us to estimate extinction rates from phylogeny

Biogeographic model

Since we know the biogeographical region (state) of the tip species in a phylogeny, we can use a Dispersal-Extinction-Cladogenesis (DEC) model to estimate the ancestral range for each node in the phylogeny
Ree et al. (2005) Evolution

Linking macroevolution and macroecology