CommPhyl

Report
Community Phylogenetic
structure with R
• Central question in community ecology
• What processes are responsible for the identity and
relative abundances of co-occurring species in local
assemblages?
• What is the relative importance of different ecological
processes in structuring communities
• Search for patterns in community structure
that may reflect underlying processes
Community
assembly
processes
Niche and
trait
evolution
Patterns of
phylogenetic
community
structure
Phylogenetic data
Patterns of phylogenetic community
structure
• Phylogenetic clustering
– Co-occurring species are
more closely related than
expected by chance
• Phylogenetic overdispersion/evenness
– Co-occurring species are
more distantly related than
expected by chance
• Random patterns
Community assembly processes
• Many processes could influence phylogenetic
community structure
– Facilitation
– Negative density dependent processes (e.g. herbivory,
predation)
– Indirect interactions mediated by herbivores,
parasites, pathogens
– Disturbance e.g. fire
– Speciation
• Two main processes are usually considered
Habitat filtering
• Adaptation of species to
abiotic conditions
– Communities
composed of
ecologically similar
species
– Environment imposes
a filter via abiotic
conditions favoring
species with similar
adaptations
Emerson & Gillepsie, 2008
Competitive exclusion
– The principle of
competitive exclusion:
similar species cannot
coexist indefinitely.
– Related species are
ecologically similar
• Prediction: closely related
species should co-occur
less than would be
expected
• Early efforts tested this
hypothesis using
– species : genus ratios
– Taxonomic community
structure
Emerson & Gillepsie, 2008
Phylogenetic signal
Kembel, 2009
Community phylogenetic structure
(Competitive exclusion)
Kembel, 2009
(or clustering)
Clustering: Co-occurring species are more closely related than expected by chance
Over-dispersion: Co-occurring species are more distantly related than expected
Clustering due to
environmental
filtering
Over-dispersion due
to competitive
exclusion
Clustering due to
competitive
exclusion
Over-dispersion due
to environmental
filtering
Trait conservatism
Trait conservatism
Trait convergence
Trait convergence
h
Emerging patterns
• Communities structured by several processes
acting in concert
– Mediated by different sets of traits
• Scale is important
– Taxonomic scale (Cavender-Barres et al., 2004; 2006)
• Oaks: phylogenetic over dispersion
• Angiosperms: phylogenetic clustering
– Spatial scales
• Similar shift from smaller (overdispersion) to larger spatial
scales (clustering)
Steps
• Quantify the degree of relatedness among cooccurring species using a phylogeny
• Define a broader pool of species from which
communities have been assembled
• Construct a null model which generates random
communities from the broader species pool
• Determine phylogenetic signal for functional
traits that influence community assembly
Data
• Phylogenetic tree for regional/broader species
pool
– Entire species in all the communities
• Species list and presence/absence or
abundance data for the different
communities, plots or specific habitats within
a community.
Getting a phylogenetic tree
• Can use either sequence data or species list
• Different formats: Newick, Nexus
• Online tools available e.g.
– Genbank (http://www.ncbi.nlm.nih.gov/): can obtain raw
sequence data
– ARB/Silva (http://www.arb-silva.de/): provides aligned sequence
data free for academic use
– RDP (http://rdp.cme.msu.edu/ ): provides sequence data and
builds phylogenetic tree
– Beast (http://beast.bio.ed.ac.uk/Main_Page): contains programs
to create phylogenies with sequence data
– Phylomatic (http://www.phylodiversity.net/phylomatic/):
assembles phylogenies using species lists.
Quantifying degree of relatedness
Vamosi et al., 2009
Within community measures
• Faith’s phylogenetic diversity (PD)
– Total branch length spanned by the tree including
all species in a local community.
• A lower value indicates that
– Taxa are clumped on the phylogeny
– Capture only a small part of the total phylogenetic
diversity present in the entire phylogeny
– Co-ocurring species are more closely related
• MPD: Mean pair-wise distance between all species in a
community
– Measures whether species in a community are more
closely related than expected by chance (using a null
model and the regional species pool)
– MPD is more sensitive to tree-wide phylogenetic patterns
• MNTD: Mean distance to nearest taxon for each
species in the community
– Measures whether closely related species tend to co-occur
or not (using a null model and the regional species pool).
– MNTD is more sensitive to patterns of evenness and
clustering closer to the tips of the phylogeny
Distance
matrix
Kembel, 2009
Null models
• Randomize the phylogeny
– Phylogeny shuffle: randomizes phylogenetic relationships
among species by shuffling the taxa on the tips.
– Randomize tree structure
• Randomize community structure
– Randomize draws from species pool
• Species in each sample are random draws from the
– Sample pool: maintains species richness of each sample but species are
drawn without replacement from the list of all species actually
occurring in the sample.
– Regional species pool: maintains species richness, but species are
drawn without replacement from a broader phylogeny pool.
– Randomize community matrix
• Independent swap: creates swapped versions of the sample/species
matrix; only applicable with presence/absence data.
Randomization
• Using null model
– Generate a random community
– Recalculate metrics i.e. MPD/MNTD
– Repeat many times
• Result: distribution of metric values for random
communities
Standardized effect sizes (SES)
• SES=(Observed value – Mean (random values))/SD(random values)
• Net relatedness index (NRI): standardized metric obtained by
comparing MPDobs and MPDexp
– + NRI = phylogenetic clustering
– - NRI = phylogenetic evenness
– Calculated as standardized effect size (SES) in picante
• SES = -1 x NRI
SES continued
• Nearest taxon index (NTI): obtained by comparing MNTDobs and
MNTDexp
– + NTI = phylogenetic clustering
– - NTI = phylogenetic evenness
– Computed as standardized effect size (SES) in picante
• SES = -1 x NTI
Phylogenetic Beta Diversity
• Measures patterns of
phylogenetic relatedness among
communities
– Among communities equivalent of
MPD and MNTD using pairs of
species drawn from different
communities
– can be used with any method
based on among-community
distances
• e.g. cluster analysis,
phyloordination, Mantel tests with
spatial/environmental distances
separating communities.
Today’s Data
• Bird communities along 3 highways: 64, 44,
and 55
– Gradient from urban (1) to rural (5) along those
highways, with each community separated by 16
km
– Ex. Communities: RD441, RD442, RD443, RD641,
RD642, etc…
1
2
3
RD64
???
4
5
• Genetic data from Genbank using 2 mitochondrial
genes
– Cytochrome b (Cytb) and Cytochrome c oxidase
subunit I (COI)
– Made a phylogeny using Beast and Beauti (free online,
easy to use, and comes with a short tutorial and good
manual)
• Question: What is the phylogenetic structure of
communities as you go from rural to urban
environments?
1
2
3
RD64
???
4
5

similar documents