Traitdependent dispersal models for phylogene4c biogeography, in the R package 'BioGeoBEARS' Nicholas J. Matzke, DECRA Fellow Moritz Lab Group, EEG, Research School of Biology The Australian NaConal University [email protected]phylo.wikidot.com/BioGeoBEARS – June 21, 2016, iEvoBio, 2:12 pm – EvoluCon 2016, AusCn, TX Adolf Meyer (1898), The Birds of Celebes and the neighbouring islands. https:// en.wikipedia.org/wiki/Snoring_rail Snoring rail, Aramidopsis plateni (flightless, Sulawesi)
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Trait-‐dependent dispersal models for phylogene4c biogeography, in the R package 'BioGeoBEARS'
Nicholas J. Matzke, DECRA Fellow Moritz Lab Group,
EEG, Research School of Biology The Australian NaConal University
-‐ ConnecCvity matrices -‐ User-‐specified dispersal probability mulCpliers -‐ Geographic distance matrices -‐ Environmental distance matrices -‐ Any of these can be “Cme-‐straCfied” -‐ Appearing/disappearing areas
Source: Kirchman, Jeremy J. (2012). Speciation of Flightless Rails on Islands: A DNA-Based Phylogeny of the Typical Rails of the Pacific, The Auk, 129(1):56-69.
Living and ex4nct Rail distribu4ons (Kirchman 2012)
Source: Kirchman, Jeremy J. (2012). Speciation of Flightless Rails on Islands: A DNA-Based Phylogeny of the Typical Rails of the Pacific, The Auk, 129(1):56-69.
Living and ex4nct Rail distribu4ons (Kirchman 2012)
Tahi4 Rail: Gallirallus pacificus
https://en.wikipedia.org/wiki/Tahiti_rail
Tahi4 Rail: Gallirallus pacificus
Source: Kirchman, Jeremy J. (2012). Speciation of Flightless Rails on Islands: A DNA-Based Phylogeny of the Typical Rails of the Pacific, The Auk, 129(1):56-69.
https://en.wikipedia.org/wiki/Tahiti_rail
Source: Kirchman, Jeremy J. (2012). Speciation of Flightless Rails on Islands: A DNA-Based Phylogeny of the Typical Rails of the Pacific, The Auk, 129(1):56-69.
Living and ex4nct Rail distribu4ons (Kirchman 2012)
Source: Kirchman, Jeremy J. (2012). Speciation of Flightless Rails on Islands: A DNA-Based Phylogeny of the Typical Rails of the Pacific, The Auk, 129(1):56-69.
Areas (simplified from Kirchman)
Source: Figure 1 of: Kirchman, Jeremy J.; Steadman, David W. (2006). New Species of Rails (Aves: Rallidae) from an Archaeological Site on Huahine, Society Islands. Pacific Science (2006), 60:2, 281–297.
Areas (simplified from Kirchman)PhillipinesSulawesiMollucasPNG, Australia
New ZealandOkinawaGuam/MarianasNew Britain
Solomon IslandsCook IslandsChatham IslandsLord Howe IslandWake Island
Dated phylogeny of typical Rails (simplified from Kirchman)
BioGeoBEARS DEC+J+t12+t21+jt12+m1 on Pacific Railsancstates: global optim, 2 areas max. d=0.0935; e=0.1565; j=0.3493; t12=9.2439; t21=3.0112; jt12=0.4168; LnL=−97.51
Clearly the group has dispersed …but probably they were flying to do it!
Build a trait-‐based dispersal model
1. Trait flight/flightlessness trait evolves on tree Parameters of model:
Clearly the group has dispersed …but probably they were flying to do it!
Build a trait-‐based dispersal model
1. Trait flight/flightlessness trait evolves on tree Parameters of model: t12 -‐> Rate of flight loss
t21 -‐> Rate of flight gain m2 -‐> MulCplier on dispersal prob. while flightless
Clearly the group has dispersed …but probably they were flying to do it!
Build a trait-‐based dispersal model
1. Trait flight/flightlessness trait evolves on tree Parameters of model: t12 -‐> Rate of flight loss
t21 -‐> Rate of flight gain m2 -‐> MulCplier on dispersal prob. while flightless
FLY NOT
FLY - t12
NOT t21 -
Clearly the group has dispersed …but probably they were flying to do it!
Build a trait-‐based dispersal model
1. Trait flight/flightlessness trait evolves on tree Parameters of model: t12 -‐> Rate of flight loss
t21 -‐> Rate of flight gain m2 -‐> MulCplier on dispersal prob. while flightless
2. Combine trait model with anageneCc dispersal
FLY NOT
FLY - t12
NOT t21 -
Clearly the group has dispersed …but probably they were flying to do it!
Build a trait-‐based dispersal model
3. CladogeneCc parameters of model: jt12 -‐> Chance of “jumping” from flight-‐> flightless during founder-‐event speciaCon
Trait-dependent DEC model
Dispersal-Extinction (on branches)
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
A
A BA B
A A B
A B
d = rate of “dispersal” (range expansion)e = rate of “extinction” (range contraction / local extirpation)
descendant rangeRate matrix:
Traditional:
Trait-dependent DEC model
Dispersal-Extinction (on branches)
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
A
A BA B
A A B
A B
d = rate of “dispersal” (range expansion)e = rate of “extinction” (range contraction / local extirpation)
descendant rangeRate matrix:
Traditional:Rate matrix:
Trait-dependent:
Trait-dependent DEC model
Dispersal-Extinction (on branches)
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
A
A BA B
A A B
A B
d = rate of “dispersal” (range expansion)e = rate of “extinction” (range contraction / local extirpation)
descendant rangeRate matrix:
Traditional:Rate matrix:
Trait-dependent:
FLY NOT
FLY - t12
NOT t21 -
Trait-dependent DEC model
Dispersal-Extinction (on branches)
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
A
A BA B
A A B
A B
d = rate of “dispersal” (range expansion)e = rate of “extinction” (range contraction / local extirpation)
descendant rangeRate matrix:
Traditional:
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
descendant range
Rate matrix:
Trait-dependent:
FLY NOT
FLY - t12
NOT t21 -
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
descendant range
FLY NOT
FLY - t12
NOT t21 -=+
trait matrix dispersal/extinction+
null A B AB
null -A e - dB e - dAB e e -
ancestorrange
descendant range
FLY NOT
FLY - t12
NOT t21 -
t21
=
null A B AB
null -A e - d
B e - d
AB e e -null A B AB
null -A e - md
B e - md
AB e e -
t12
+
Combined anagenetic
matrix
trait matrix dispersal/extinction+
DEC+J model: add trait to jump (Matzke 2014)
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
A
A BA B
A A B
A BCladogenesis(at splits)
Equalper-event
weights
(modified by j in DEC+J)
Jump dispersal (founder events)
Traditional:
DEC+J model: add trait to jump (Matzke 2014)
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
A
A BA B
A A B
A BCladogenesis(at splits)
Equalper-event
weights
(modified by j in DEC+J)
Cladogenesis(at splits)
Jump dispersal (founder events)
Traditional: Trait-dependent:
DEC+J model: add trait to jump (Matzke 2014)
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
A
A BA B
A A B
A BCladogenesis(at splits)
Equalper-event
weights
(modified by j in DEC+J)
Cladogenesis(at splits)
Jump dispersal (founder events)
Traditional: Trait-dependent:
DEC+J model: add trait to jump (Matzke 2014)
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
A
A BA B
A A B
A BCladogenesis(at splits)
Equalper-event
weights
(modified by j in DEC+J)
Cladogenesis(at splits)
Jump dispersal (founder events)
Trait model:
Traditional: Trait-dependent:
DEC+J model: add trait to jump (Matzke 2014)
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
A
A BA B
A A B
A BCladogenesis(at splits)
Equalper-event
weights
(modified by j in DEC+J)
Cladogenesis(at splits)
Jump dispersal (founder events)
m2 xTrait model:
Traditional: Trait-dependent:
Cladogenetic range-change model: DEC
Cladogenetic range-change events:
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
ending&rangeleft: A A A A A A A B B B B B B B C C C C C C C AB AB AB AB AB AB AB BC BC BC BC BC BC BC AC AC AC AC AC AC AC ABCABCABCABCABCABCABCright: A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC countA sym 1B sym 1
starting C sym 1range AB vic sub vic sub sub sub 6
BC vic sub vic sub sub sub 6AC vic sub vic sub sub sub 6ABC vic sub vic sub vic sub vic vic vic sub sub sub 12
jjjj
jjj
jj j
jj j
j
jj j
j
Cladogenetic transition matrix:
ending&rangeleft: A A A A A A A B B B B B B B C C C C C C C AB AB AB AB AB AB AB BC BC BC BC BC BC BC AC AC AC AC AC AC AC ABCABCABCABCABCABCABCright: A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC countA sym 1B sym 1
starting C sym 1range AB vic sub vic sub sub sub 6
BC vic sub vic sub sub sub 6AC vic sub vic sub sub sub 6ABC vic sub vic sub vic sub vic vic vic sub sub sub 12
jjjj
jjj
jj j
jj j
5 5 5 8 8 8 12
j
jj j
j
Cladogenetic transition matrix:
Cladogenetic range-change events:
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
Founder-event speciation
Cladogenetic range-change model: DEC+J
ending&rangeleft: A A A A A A A B B B B B B B C C C C C C C AB AB AB AB AB AB AB BC BC BC BC BC BC BC AC AC AC AC AC AC AC ABCABCABCABCABCABCABCright: A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC countA sym 1B sym 1
starting C sym 1range AB vic sub vic sub sub sub 6
BC vic sub vic sub sub sub 6AC vic sub vic sub sub sub 6ABC vic sub vic sub vic sub vic vic vic sub sub sub 12
jjjj
jjj
jj j
jj j
5 5 5 8 8 8 12
j
jj j
j
Cladogenetic transition matrix while flying:
Cladogenetic range-change events:
Sympatry (narrow)
Sympatry (subset)
Vicariance (narrow)
Founder-event speciation
Trait-based cladogenetic model: DEC+J
ending&rangeleft: A A A A A A A B B B B B B B C C C C C C C AB AB AB AB AB AB AB BC BC BC BC BC BC BC AC AC AC AC AC AC AC ABCABCABCABCABCABCABCright: A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC A B C AB BC AC ABC countA sym 1B sym 1
starting C sym 1range AB vic sub vic sub sub sub 6
BC vic sub vic sub sub sub 6AC vic sub vic sub sub sub 6ABC vic sub vic sub vic sub vic vic vic sub sub sub 12
mjj 5
5 5 8 8 8 12
j
Cladogenetic transition matrix while non-flying: j = m x j
BioGeoBEARS DEC+J+t12+t21+jt12+m1 on Pacific Rail flightancstates: global optim, 1 areas max. d=0.0935; e=0.1565; j=0.3493; t12=9.2439; t21=3.0112; jt12=0.4168; LnL=−97.51
Not
NotFly
Fly
Fly
Fly
Fly
Fly
Fly
Fly
Fly
Fly
Fly
FlyFly
FlyNot
FlyFly
FlyFly
FlyNot
NotNot
Not
Not
Fly
Fly
Fly
Fly
Fly
Fly
Fly
FlyFlyFly
Fly
FlyFly
FlyNot
FlyFly
FlyFly
FlyNot
NotNot
1. Trait-‐based dispersal models are feasible, at least for moderate-‐sized problems (key issue: size of matrices)
2. Trait evo. & its influence on dispersal can be revealed by parameter es@ma@on 3. Combinable with distance, etc. 4. Adding to BioGeoBEARS; email me to try it: [email protected]
Conclusions
Adolf Meyer (1898), The Birds of Celebes and the neighbouring islands. https://