networks and species delimitations

Using the multispecies coalescent model to infer species tree/networks and species delimitations Salvatore Tomasello, Florian Wagner, Christoph Oberp...
Author: Shannon Sims
9 downloads 0 Views 3MB Size
Using the multispecies coalescent model to infer species tree/networks and species delimitations

Salvatore Tomasello, Florian Wagner, Christoph Oberprieler

07/06/2016

Introduction

https://commons.wikimedia.org/wiki/File:Asteracea_poster_3.jpg

Introduction

Species delimitation: “the process of determining the boundaries and numbers of species from empirical data”

So far being done with: ⁻ ⁻ ⁻ ⁻ ⁻

Subjective quantification of morphological variation Crossing experiments Phenology Secondary compounds Testing monophyly in phylogenies

„It belongs to the wise man to order“ (Thomas Aquinas)

Introduction

Phylogenetic species concept: “a species is an irreducible cluster of organisms, diagnosably distinct from other such clusters, and within which there is a parental pattern of ancestry and descendent” (Cracraft, 1989)

- Reciprocal monophyly

Introduction

Species-level phylogenetics • multi-locus sequence data

Introduction

The coalescent Theory “The mathematical and probabilistic theory underlying the evolutionary history of alleles” The probability of two alleles of coalesce back in time is function of: - Time (τ) - Population size (θ)

From: Fujita et al. 2012. Coalescent-based species delimitation in an integrative taxonomy. Trends in Ecology and Evolution

Introduction Species delimitation based on multi-locus sequence • coalescent theory allows a probabilistic approach to species delimitation • accommodates gene tree discordance • does not require reciprocal monophyly of alleles or fixed differences • provides framework for determining depth (divergence times) and width (effective population sizes) of species trees • objective testing of alternative hypotheses of evolutionary independence using multilocus data (sequences, fingerprints)

Introduction When are molecular approaches for species delimitation useful • • • •

Absence of taxonomically relevant characters Unexplored areas with high diversity Allopatric populations (species?) Cryptic speciation

„monspeliense“

„ligusticum“

„vulgare“ „ageratifolium“

„legraeanum“

„laciniatum“

Method based on the coalescent:

GMYC

GMYC

The General Mixed Yule Coalescent method (GMYC) • identifies the transition points between inter- and intra-species branching rates • Maximum likelihood method (ML) • Based on the Yule process • Single locus data • Ultrametric tree as input • Web server: http://species.h-its.org/gmyc/

• Available at: https://sites.google.com/site/noahmreid/software

Pons et al. 2005. Sequence-Based Species Delimitation for the DANN Taxonomy of Undescribed Insects. – Systematic Biology 55: 595-609. Reid & Carstens. 2012. Phylogenetic estimation error can decrease the accuracy of species delimitation: a Bayesian implementation of the general mixed Yule-coalescent model. – BMC Evolutionary Biology 12:196

GMYC

Method based on the coalescent:

GMYC

The genus Orychophragmus • Five species before recognised • North, Central and S-Eastern China

• O. violaceous widespread polymorphic, with several infraspecific taxa

• 160 individuals (63 pop.) • Four plastid markers (matK, psbA-trnH, rbcL, trnL-F) • Non coalescent-based method: Phylogenetic species concept http://commons.wikimedia. org/wiki/User:KENPEI

• Non coalescent-based method: TCS

• Coalescent-based method: GMYC Hu et al. 2015. Species delimitation in Orychophragmus (Brassicaceae) based on chloroplast and nuclear DNA barcodes. – Taxon 64:714-726.

Method based on the coalescent:

GMYC

GMYC

• 8 clusters identified by TCS • 11 species identified by GMYC • Two new species recognised in O. violaceous (P3, P4) • The new taxa are also identifiable with morphological traits

Hu et al. 2015. Species delimitation in Orychophragmus (Brassicaceae) based on chloroplast and nuclear DNA barcodes. – Taxon 64:714-726.

Method based on the coalescent:

GMYC

GMYC

The General Mixed Yule Coalescent method (GMYC) • identifies the transition points between inter- and intra-species branching rates • Maximum likelihood method (ML) • Based on the Yule process • Single locus data • Ultrametric tree as input • Web server: http://species.h-its.org/gmyc/

• Available at: https://sites.google.com/site/noahmreid/software Relatively fast for large dataset

Depende on the accurancy of the ultrametric tree

Pons et al. 2005. Sequence-Based Species Delimitation for the DANN Taxonomy of Undescribed Insects. – Systematic Biology 55: 595-609. Reid & Carstens. 2012. Phylogenetic estimation error can decrease the accuracy of species delimitation: a Bayesian implementation of the general mixed Yule-coalescent model. – BMC Evolutionary Biology 12:196

SpedeSTEM

SpedeSTEM • Based of the species tree reconstruction method STEM • Maximum likelihood method (ML) • Use sequence information as input • Multiple loci • Likelihood ratio test or AIC to find the best species delimitation model • Available at: http://carstenslab.org.ohio-state.edu/software.html

Ence & Carstens. 2011. SpedeSTEM: a rapid and accurate method for species delimitation. – Molecular Ecology Resources 11: 473-480.

spedeSTEM SpedeSTEM

Method based on the coalescent:

The genus Protoparmelia (Lecanorales, Ascomycota) • Morphologically heterogeneous genus • Polyphyletic

• Very diverse ecological requirements: boreal-arctic/alpine, temperate, Mediterranean, subtropical, and tropical species • P. badia, P. memnonia cosmopolitan species • Tropical regions not very explored

lichensmaritimes.org

– 160 Individuals from 18 species (70%) – 6 markers

P. montagnei

– Phylogeny of the genus

P. badia stridvall.se

– SpedeSTEM – BP&P P. badia

Singh et al. 2015. Coalescent-Based Species Delimitation Approach Uncovers High Cryptic Diversity in the Cosmopolitan Lichen-Forming Fungal Genus Protoparmelia (Lecanorales, Ascomycota). – PLoS ONE 10: e0124625. doi:10.1371/journal.pone.0124625

Method based on the coalescent:

spedeSTEM SpedeSTEM

The genus Protoparmelia s.l. • Two clades within Protoparmelia s. str. • One with temperate species

• The other with tropical species • Six species of Protoparmelia claster with the genus Miriquidica

Singh et al. 2015. Coalescent-Based Species Delimitation Approach Uncovers High Cryptic Diversity in the Cosmopolitan Lichen-Forming Fungal Genus Protoparmelia (Lecanorales, Ascomycota). – PLoS ONE 10: e0124625. doi:10.1371/journal.pone.0124625

Method based on the coalescent:

spedeSTEM SpedeSTEM

The genus Protoparmelia s.str. • Well supported clades as putative species • 25 species-scenario vs. the 12 species

previously recognised • SpedeSTEM supported 19 species • BP&P supported a 23-species model

Singh et al. 2015. Coalescent-Based Species Delimitation Approach Uncovers High Cryptic Diversity in the Cosmopolitan Lichen-Forming Fungal Genus Protoparmelia (Lecanorales, Ascomycota). – PLoS ONE 10: e0124625. doi:10.1371/journal.pone.0124625

Method based on the coalescent:

BP&P

BP&P

Bayesian Phylogenetics and Phylogeography (BP&P) • Bayesian modeling approach generating speciation posterior probabilities • Sequence data as input • Multiple loci • Uncertainty in gene phylogeny • Guide species tree as input • Prior for θ: effective population size parameter

• Prior for τ: species divergence time parameter • Available at: http://abacus.gene.ucl.ac.uk/software.html

Rannala & Yang. 2003. Bayes Estimation of Species Divergence Times and Ancestral Population Sizes Using DNA Sequences From Multiple Loci. – Genetics 164: 1645-1656. Yang & Rannala. 2010. Bayesian species delimitation using multilocus sequence data. – PNAS 107: 9264–9269.

Method based on the coalescent:

BP&P

BP&P

Species Delimitation in Otatea (Poaceae, Guaduinae) • Eight accepted species • Seven endemic of Mexico

• Otatea fimbriata also distributed in central America and Colombia • O. acuminata widespread species growing in tropical dry forest of the Pacific coast

Ruiz.Sanchez. 2015. Parametric and non-parametric species delimitation methods result in the recognition of two new Neotropical woody bamboo species. – Molecular Phylogenetics and Evolution 93: 261-273

BP&P

Method based on the coalescent:

BP&P

Species Delimitation in Otatea (Poaceae, Guaduinae) • Eight accepted species • Seven endemic of Mexico

• Otatea fimbriata also distributed in central America and Colombia • O. acuminata widespread species growing in tropical dry forest of the Pacific coast • Two diverging populations of O. acuminata (Chapas, Hidalgo) • Three plastid regions • nrDNA ITS • Morphometrical analyses • BP&P • Different sets of priors for θ, τ

Ruiz.Sanchez. 2015. Parametric and non-parametric species delimitation methods result in the recognition of two new Neotropical woody bamboo species. – Molecular Phylogenetics and Evolution 93: 261-273

Method based on the coalescent:

BP&P

BP&P

• All species monophyletic • O. acuminata polyphyletic

Ruiz.Sanchez. 2015. Parametric and non-parametric species delimitation methods result in the recognition of two new Neotropical woody bamboo species. – Molecular Phylogenetics and Evolution 93: 261-273

Method based on the coalescent: Concatenated tree

BP&P

BP&P

BP&P result

Ruiz.Sanchez. 2015. Parametric and non-parametric species delimitation methods result in the recognition of two new Neotropical woody bamboo species. – Molecular Phylogenetics and Evolution 93: 261-273

Method based on the coalescent:

BP&P

BP&P

Bayesian Phylogenetics and Phylogeography (BP&P) • Bayesian modeling approach generating speciation posterior probabilities • Sequence data as input • Multiple loci • Uncertainty in gene phylogeny • Without guide species tree as input (simultaneous species delimitation and species tree inference) • Prior for θ: effective population size parameter • Prior for τ: species divergence time parameter • Available at: http://abacus.gene.ucl.ac.uk/software.html

Yang & Rannala. 2014. Unguided Species Delimitation Using DNA Sequence Data from Multiple Loci – Mol. Biol. Evol. 31: 3125-3135.

Using marginal likelihood estimations Using marginal Likelihood

Method based on the coalescent:

Marginal Likelihood for species delimitation

Bayes‘ formula:

prior probability

likelihood

𝑃 𝐻𝑎 ∙𝑃 𝐷 𝐻𝑎 𝑃 𝐻𝑎 𝐷 = 𝑖 𝑃 𝐻𝑖 ∙𝑃 𝐷 𝐻𝑖 posterior probability

marginal likelihood

Method based on the coalescent:

𝑖

𝑃 𝐻𝑖 ∙ 𝑃 𝐷 𝐻𝑖

Using marginal Likelihood

𝑖

𝑃 𝐻𝑖 ∙ 𝑃 𝐷 𝐻𝑖

C

C

B

B A A 3 species model

2 species model

Using marginal Likelihood

• Silene sect. Cryptoneurae Aydin & Oxelman:  Silene insularis Barbey  S. salamandra Pamp.  S. cryptoneura Stapf  S. ertekinii Aydin & Oxelman • current species delimitations mainly based on floral characteristics & geographical distribution

Aydin et al. 2014. Marginal Likelihood Estimate Comparisons to Obtain Optimal Species Delimitations in Silene sect. Cryptoneurae (Caryophyllaceae). – PLoS ONE 9: e106990. doi:10.1371/journal.pone.0106990

Using marginal Likelihood

S. ertekinii (E) S. salamandra (S)

S. cryptoneura (W)

Silene insularis (I) Map of Southwest Anatolia and Aegean Islands of Rhodes and Karpathos showing the known geographic distributions of Silene sect. Cryptoneurae. Aydin et al. 2014. Marginal Likelihood Estimate Comparisons to Obtain Optimal Species Delimitations in Silene sect. Cryptoneurae (Caryophyllaceae). – PLoS ONE 9: e106990. doi:10.1371/journal.pone.0106990

Using marginal Likelihood

species delimitation in sect. cryptoneurae:  S. ertekinii & S. cryptoneura are different species

Model

species del.

1

EWSI

1

2

EW+S+I

3

3

E+W+S+I

4

4

E+W+SI

3

5

E+WSI

2

6

ESI+W

2

7

EW+SI

2

8

EWI+S

2

9

EWS+I

2

S. ertekinii (E) S. salamandra (S)

S. cryptoneura (W)

Silene insularis (I)

Aydin et al. 2014. Marginal Likelihood Estimate Comparisons to Obtain Optimal Species Delimitations in Silene sect. Cryptoneurae (Caryophyllaceae). – PLoS ONE 9: e106990. doi:10.1371/journal.pone.0106990

# species

Probability of gene trees

The tetraploid Iberian Leucanthemopsis:

Actual taxonomic treatment ⁻

L(MUL-tree)



AIC = -2 lnL + 2k

L. flaveola L. pallida var. alpina

L. pallida subsp. spatulifolia L. pallida subsp. pallida

Probability of gene trees

The tetraploid Iberian Leucanthemopsis:

Best taxa delimitation scenario ⁻

lnL = 2,996.31



AIC = 6,026.62

L. flaveola L. pallida var. alpina

L. pallida subsp. spatulifolia L. pallida subsp. pallida

Method based on the coalescent:

BP&P

Bayesian species delimitation using molecular and morphological data (iBP&P) • Quantitative and continuous traits • Bayesian inference based on a joined analysis of molecular and morphological data

Integrative approach!!

Solis-Lamus et al. 2014. Bayesian species delimitation combining multiple genes and traits in a unified framework. – Evolution 62: 492-507

Problems of species delimitation: • Hybridization (Allopolyploidy)

„Systematics is the stone of Sisyphus“ (Vernon Heywood)

„One should imagine Sisyphus happy“ (Albert Camus)

Thanks for the attention…