This document summarizes a presentation on a study that used full Bayesian comparative methods and RADseq genomic data to analyze the phylogeography of two genera of geckos in the Philippines, Cyrtodactylus and Gekko. The study found strong support for independent divergences of Cyrtodactylus populations across islands but weaker support for independent divergences in Gekko. This provides evidence against the "species pump" hypothesis that fragmentation of islands alone promoted diversification in these lizards. However, the study notes limitations including few island pairs analyzed and variation in fragmentation times among islands.
Full Bayesian comparative biogeography of Philippine geckos challenges predictions of climate-driven vicariant speciation
1. Full Bayesian comparative biogeography of Philippine geckos
challenges predictions of climate-driven vicariant speciation
Evolution 2019
Jamie Oaks
Auburn University
Web: phyletica.org
GitHub: joaks1 & phyletica
Twitter: @jamoaks
Slides: phyletica.org/slides/evol2019.pdf
Cameron Siler Rafe Brown
University of Oklahoma University of Kansas
Thanks to ASN, SSE, SSB, and all organizers!
Comparative biogeography of geckos @jamoaks 1/16
4. Did fragmentation of islands
promote diversification?
Comparative biogeography of geckos @jamoaks 3/16
5. Click here for a sea-level
animation of SE Asia
youtu.be/mLNLRdbu5W8
Did fragmentation of islands
promote diversification?
Comparative biogeography of geckos @jamoaks 3/16
9. m1 m2 m3 m4 m5
We want to infer the model and divergence times given genetic data
Comparative biogeography of geckos @jamoaks 5/16
10. Previous tests of “species-pump”
Our previous attempts using ABC left
little information in summary statistics
to inform divergence times 1,2
0.00
0.05
0.10
0.15
0.20
1 2 3 4 5 6 7 8 9
Number of events
Probability
Insulasaurus
Ptenochirus
Macroglossus
Haplonycteris
Gekko
Dendrolaphis
Cyrtodactylus
Cynopterus
Crocidura
0 5 10 15
Time
Comparison
1
J. R. Oaks et al. (2013). Evolution 67: 991–1010. 2
J. R. Oaks (2014). BMC Evolutionary Biology 14: 150
Comparative biogeography of geckos @jamoaks 6/16
11. Previous tests of “species-pump”
Our previous attempts using ABC left
little information in summary statistics
to inform divergence times 1,2
What now? 0.00
0.05
0.10
0.15
0.20
1 2 3 4 5 6 7 8 9
Number of events
Probability
Insulasaurus
Ptenochirus
Macroglossus
Haplonycteris
Gekko
Dendrolaphis
Cyrtodactylus
Cynopterus
Crocidura
0 5 10 15
Time
Comparison
1
J. R. Oaks et al. (2013). Evolution 67: 991–1010. 2
J. R. Oaks (2014). BMC Evolutionary Biology 14: 150
Comparative biogeography of geckos @jamoaks 6/16
12. Previous tests of “species-pump”
Our previous attempts using ABC left
little information in summary statistics
to inform divergence times 1,2
What now?
We need more data and/or a more
efficient method
0.00
0.05
0.10
0.15
0.20
1 2 3 4 5 6 7 8 9
Number of events
Probability
Insulasaurus
Ptenochirus
Macroglossus
Haplonycteris
Gekko
Dendrolaphis
Cyrtodactylus
Cynopterus
Crocidura
0 5 10 15
Time
Comparison
1
J. R. Oaks et al. (2013). Evolution 67: 991–1010. 2
J. R. Oaks (2014). BMC Evolutionary Biology 14: 150
Comparative biogeography of geckos @jamoaks 6/16
13. Previous tests of “species-pump”
Our previous attempts using ABC left
little information in summary statistics
to inform divergence times 1,2
What now?
We need more data and/or a more
efficient method
Our goal is to do both
0.00
0.05
0.10
0.15
0.20
1 2 3 4 5 6 7 8 9
Number of events
Probability
Insulasaurus
Ptenochirus
Macroglossus
Haplonycteris
Gekko
Dendrolaphis
Cyrtodactylus
Cynopterus
Crocidura
0 5 10 15
Time
Comparison
1
J. R. Oaks et al. (2013). Evolution 67: 991–1010. 2
J. R. Oaks (2014). BMC Evolutionary Biology 14: 150
Comparative biogeography of geckos @jamoaks 6/16
14. Cyrtodactylus Gekko
Latitude q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
Bohol
Palawan
Samar
Luzon 1
Luzon 2
Polillo
Panay
Sibuyan
Camiguin Sur
Borneo
Leyte
Babuyan Claro
Camiguin Norte
Luzon 3
Negros
Tablas
Connected?
Yes
Maybe
No
10
15
20
118 120 122 124 126
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
Babuyan Claro
North Gigante
Lubang
Masbate
Negros
Sabtang
Romblon
Camiguin Norte
Calayan
South Gigante
Luzon
Panay 1
Panay 2
Batan
Tablas
Dalupiri
10
15
20
118 120 122 124 126
Longitude
Sampled 2–5 individuals from 8 pairs of populations for both Cyrtodactylus and Gekko
Collected short DNA sequences (RADseq) from across genome of each individual
J. R. Oaks et al. (2019). Evolution 73: 1151–1167
Comparative biogeography of geckos @jamoaks 7/16
15. Analyzed RADseq data with full-likelihood Bayesian comparative phylogeographic method1:
Ecoevolity: Estimating evolutionary coevality
1
J. R. Oaks (2019). Systematic Biology 68: 371–395
2
D. Bryant et al. (2012). Molecular Biology and Evolution 29: 1917–1932
Comparative biogeography of geckos @jamoaks 8/16
16. Analyzed RADseq data with full-likelihood Bayesian comparative phylogeographic method1:
Ecoevolity: Estimating evolutionary coevality
Analytically integrate over coalescent genealogies and mutational histories2
Dirichlet-process prior across divergence models
1
J. R. Oaks (2019). Systematic Biology 68: 371–395
2
D. Bryant et al. (2012). Molecular Biology and Evolution 29: 1917–1932
Comparative biogeography of geckos @jamoaks 8/16
17. Analyzed RADseq data with full-likelihood Bayesian comparative phylogeographic method1:
Ecoevolity: Estimating evolutionary coevality
Analytically integrate over coalescent genealogies and mutational histories2
Dirichlet-process prior across divergence models
Goal: Fast, full-likelihood Bayesian method to infer patterns of co-diversification from
genome-scale data
1
J. R. Oaks (2019). Systematic Biology 68: 371–395
2
D. Bryant et al. (2012). Molecular Biology and Evolution 29: 1917–1932
Comparative biogeography of geckos @jamoaks 8/16
18. Analyzed RADseq data with full-likelihood Bayesian comparative phylogeographic method1:
Ecoevolity: Estimating evolutionary coevality
Analytically integrate over coalescent genealogies and mutational histories2
Dirichlet-process prior across divergence models
Goal: Fast, full-likelihood Bayesian method to infer patterns of co-diversification from
genome-scale data
Used simulations to assess how well ecoevolity works given the gekkonid RADseq data
1
J. R. Oaks (2019). Systematic Biology 68: 371–395
2
D. Bryant et al. (2012). Molecular Biology and Evolution 29: 1917–1932
Comparative biogeography of geckos @jamoaks 8/16
19. Analyzed RADseq data with full-likelihood Bayesian comparative phylogeographic method1:
Ecoevolity: Estimating evolutionary coevality
Want to learn more?
Check out Kerry Cobb’s poster (Saturday at 5:30pm ExHallBC 18)
I’ll be at the software bazaar (Sunday at 6:30pm)
1
J. R. Oaks (2019). Systematic Biology 68: 371–395
2
D. Bryant et al. (2012). Molecular Biology and Evolution 29: 1917–1932
Comparative biogeography of geckos @jamoaks 8/16
21. Results: Simulations
J. R. Oaks et al. (2019). Evolution 73: 1151–1167
Comparative biogeography of geckos @jamoaks 9/16
22. Results: Simulations
Method performs well on simulated data
J. R. Oaks et al. (2019). Evolution 73: 1151–1167
Comparative biogeography of geckos @jamoaks 9/16
27. Results: Gekko
<0.0131
<0.0075
<0.00439
<0.00245
<0.00124
0.6
1.65
1.21
0.0
0.2
0.4
0.6
1 2 3 4 5 6 7 8
Number of events
Probability
Camiguin Norte | Dalupiri
Romblon | Tablas
Sabtang | Batan
Negros | Panay 2
Panay 1 | Masbate
Lubang | Luzon
S. Gigante | N. Gigante
Babuyan Claro | Calayan
0e+00 2e−04 4e−04 6e−04 8e−04
Divergence time (subsitutions/site)
Islandpair
J. R. Oaks et al. (2019). Evolution 73: 1151–1167
Comparative biogeography of geckos @jamoaks 11/16
28. Results: Gekko
<0.0131
<0.0075
<0.00439
<0.00245
<0.00124
0.6
1.65
1.21
0.0
0.2
0.4
0.6
1 2 3 4 5 6 7 8
Number of events
Probability
Camiguin Norte | Dalupiri
Romblon | Tablas
Sabtang | Batan
Negros | Panay 2
Panay 1 | Masbate
Lubang | Luzon
S. Gigante | N. Gigante
Babuyan Claro | Calayan
0e+00 2e−04 4e−04 6e−04 8e−04
Divergence time (subsitutions/site)
Islandpair
Weak support for independent divergences
J. R. Oaks et al. (2019). Evolution 73: 1151–1167
Comparative biogeography of geckos @jamoaks 11/16
29. Take home points
Support against the “species-pump” hypothesis
Comparative biogeography of geckos @jamoaks 12/16
30. Take home points
Support against the “species-pump” hypothesis
Habitat heterogeniety and rare over-water dispersal via rafting on vegetation are likely
more important
Comparative biogeography of geckos @jamoaks 12/16
31. Caveats
Too few island pairs to rule out climate-driven vicariant speciation
Variation in fragmentation times among island pairs
Variation in mutation rates
Comparative biogeography of geckos @jamoaks 13/16
32. Caveats
Too few island pairs to rule out climate-driven vicariant speciation
Variation in fragmentation times among island pairs
Variation in mutation rates
Seems safe to conclude that the “species-pump” is not the rule for Philippine gekkonids
Comparative biogeography of geckos @jamoaks 13/16
33. Everything is on GitHub. . .
Software:
Ecoevolity: https://github.com/phyletica/ecoevolity
Open-Science Notebooks:
Ecoevolity testing: https://github.com/phyletica/ecoevolity-experiments
Gecko RADseq: https://github.com/phyletica/gekgo
Comparative biogeography of geckos @jamoaks 14/16
34. Acknowledgments
Ideas and feedback:
Phyletica Lab (the Phyleticians)
David Bryant, Mark Holder, Adam Leach´e,
and Vladimir Minin; Editors Laura
Kubatko, Mohamed Noor, and David
Weisrock; and five anonymous
Lab work:
Patrick Monnahan and John Kelly for their
help with the MSG libraries
Computation:
Alabama Supercomputer Authority
Auburn University Hopper Cluster
Funding:
DEB 1656004
Photo credits:
Rafe Brown and Cam Siler
PhyloPic!
Thanks to ASN, SSE, SSB, and all organizers of Evolution 2019!
Comparative biogeography of geckos @jamoaks 15/16