the dynamic discipline of species 2 delimitation: progress ... · paramount for conservation, human...

2The Dynamic Discipline of SpeciesDelimitation Progress TowardEffectively Recognizing SpeciesBoundaries in Natural PopulationsSteven D Leavitt Corrie S Moreau and H Thorsten Lumbsch

Contents

21 Introduction Whatrsquos in a Name TheImportance of Accurate SpeciesDelimitations 12

211 Species Concepts and Criteria 12212 Species in Lichenized Fungi Cases

of Cryptic Diversity Polymorphic Lineagesand Striking Biogeographic Patterns 13

22 A Practical Guide to ContemporarySpecies Delimitation 16

221 Corroborating Traditional Taxonomyand Discovering Cryptic Species UsingSingle-Locus Data 17

222 Sampling Across the Genome MultilocusSequence Data and Genome-Wide Markers 26

23 Can We Make Species Delimitationin Lichen-Forming Fungi TrulyIntegrative 30

231 Selecting the Appropriate Data 32

24 Conclusions What About Taxonomy 34

References 35

Abstract

Species represent a fundamental unit in evolu-tionary biology and provide a valuable contextfor organizing evaluating and communicatingimportant biological concepts and principlesEmpirical species delimitation is a dynamicdiscipline with ongoing methodological andbioinformatical developments Novel analyticalmethods increasing availability of geneticgeno-mic data increasing computational power reas-sessments of morphological and chemicalcharacters and improved availability of distribu-tional and ecological records offer excitingavenues for empirically exploring species delim-itation and evolutionary relationships amongspecies-level lineages In this chapter we aimto contribute a contemporary perspective ondelimiting species including a brief discussionon species concepts and practical direction forempirical species delimitation studies Usinglichen-forming fungi as an example we illustratethe importance and difficulties in documentingand describing species-level biodiversity

Keywords

Barcoding Coalescence DNA taxonomy Fungi Gene tree Genomics Lichens Species circumscription Species concept Species treeSD Leavitt (amp)

Committee on Evolutionary BiologyUniversity of Chicago Chicago IL USAe-mail sleavittfieldmuseumorg

SD Leavitt CS Moreau H Thorsten LumbschDepartment of Science and Education Field Museumof Natural History Chicago IL USA

copy Springer India 2015DK Upreti et al (eds) Recent Advances in LichenologyDOI 101007978-81-322-2235-4_2

11

21 Introduction Whatrsquosin a Name The Importanceof Accurate SpeciesDelimitations

Although there are over 15 million species for-mally named by scientists current estimates ofthe number of species alive on the planet todayrange from approximately two million to overone hundred million (Caley et al 2014) Docu-menting describing and naming this diversity isparamount for conservation human health foodsecurity and recreation (Tewksbury et al 2014)In a broad sense species delimitation is theprocess of identifying how individuals andpopulations fit into natural species-level clustersand not simply constructs of classification(Carstens et al 2013) Empirical species delimi-tation is a dynamic discipline with ongoingmethodological and bioinformatical develop-ments due to a growing interest in empiricalspecies delimitations Novel analytical methodsincreasing availability of geneticgenomic dataincreasing computation power reassessments ofmorphological and chemical characters andimproved availability of distributional and eco-logical records offer exciting avenues for empir-ically exploring species delimitation andevolutionary relationships in species all over theworld In this chapter we aim to contribute acontemporary perspective on delimiting speciesand offer practical direction for empirical speciesdelimitation studies

To illustrate the importance and difficulties indocumenting and describing biodiversity we willuse the lichens as an example Lichens describe amutualism between a fungus (mycobiont) and aphotosynthetic partner (photobiont) which can beeither a green algae andor cyanobacteriumLichens are ubiquitous components of most ter-restrial ecosystems playing important ecologicalroles and contributing to global biogeochemicalcycles (Porada et al 2014 Bonan and Shugart1989 Longton 1997) Due to the fact that manylichens live and grow continuously for decades oreven hundreds of years showing cumulativeresponses to changes in atmospheric pollutionlevels landmanagement practices and fluctuation

climate these are commonly used as bioindicatorsto monitor the impacts of air pollution forest agesoil quality and climate change (McCune 2000)As iconic examples of symbiosis lichens alsoprovide crucial insight into general patterns andprocesses in symbiotic systems Central tounderstanding the dynamic roles of lichens is ourability to accurately delimit and recognize speciesboundaries Increased accuracy in recognizingspecies boundaries in lichenized fungi has majorimplications for enhancing our perspective onbiological diversity evolution ecology symbi-otic interactions biomonitoring research andconservation policy

211 Species Concepts and Criteria

Species serve as a central unit for categorizingbiological diversity Humans including bird-watchers doctors fisherman gardeners politi-cians scientists and others rely to varyingdegrees on recognizing species for distinguishingdifferent kinds of organisms and effective com-munication In the biological sciences speciesrepresent one of the most fundamental units andprovide a valuable context for organizing eval-uating and communicating important biologicalconcepts and principles (Coyne and Orr 2004Mayr 1963 Darwin 1859) Due to the fact thatbiological information is commonly providedwith reference to a species unit accurate speciescircumscriptions are integral to interpreting bio-logical patterns and processes across a widerange of subdisciplines in biology (eg anatomybehavior ecology evolution physiology etc)

In spite of the underlying importance of spe-cies in biology the conceptualization of the termldquospeciesrdquo remains somewhat contentious (deQueiroz 2007 Hausdorf 2011 Coyne and Orr2004 Mayden 1997 Simpson 1951 Mayr1963) Most biologists agree that biological nat-ure is aggregated into discrete evolutionarilyindependent entities ie ldquospeciesrdquo (Coyne andOrr 2004) although theorists and empiricistsalike continue to debate over an all-encompass-ing species concept and appropriate operationalcriteria for delimiting species (Hausdorf 2011

12 SD Leavitt et al

de Queiroz 2007 Hey 2006 Donoghue andGauthier 2004 Cracraft 1983 Mishler andBrandon 1987 Mayr 1970) Over two-dozendifferent and at least partially incompatible spe-cies concepts have been proposed each based ondistinct biological properties eg differences ingenetic or morphological features adaptivezones or ecological niches mate recognitionsystems reproductive compatibility monophylyetc (de Queiroz 2007 Mayden 1997) Hausdorf(2011) argues that most species concepts areuseable but acceptance of a specific conceptrequires an appropriate adaptation of the termldquospeciesrdquo and of species delimitation In contrastde Queiroz (1998) and Mayden (1997) argue thatdistinct species concepts emphasize differentproperties of species rather than fundamentalconceptual differences and all modern speciesconcepts share an important commonalitymdashequating species with segments of metapopula-tion lineages This ldquogeneral lineage conceptrdquo(GLC de Queiroz 1999) highlights that no singleproperty should be regarded as defining for therecognition of species apart from forming lin-eages (Simpson 1951) and segments of meta-population lineages (ie ldquospeciesrdquo) may existregardless of our ability to empirically delimitthem (Camargo and Sites 2013)

We concur that the GLC provides a practicalsolution to the species concept impasse and ourdiscussion of species delimitation is based on theGLC Arguably the major implication of the GLCis that most of the earlier species concepts shouldbe regarded as secondary species ldquocriteriardquo ratherthan ldquoconceptsrdquo that can provide evidence oflineages separation (Sites and Marshall 2003Camargo and Sites 2013 Mayden 1997 de Que-iroz 2007) This pivotal distinction disentanglesthe conceptual issues of defining ldquospeciesrdquo frommethodological issues of delimiting speciesboundaries (Camargo and Sites 2013) The GLCallows researchers to delimit species using differ-ent empirical properties and facilitates the devel-opment of new methods to test hypotheses oflineage separation (de Queiroz 2007) Althoughdifferent datasets and operational criteria may giveconflicting or ambiguous results due to multipleevolutionary processes occurring within and

between populations (eg Miralles and Vences2013 Satler et al 2013) the use of several inde-pendent suites of characters such as genetic datamorphology geographic range host preferencechemistry and cross-validation using inferencesfrom multiple empirical operational criteria canprovide robust hypotheses of species boundaries(Carstens et al 2013 Fujita et al 2012)

212 Species in Lichenized FungiCases of Cryptic DiversityPolymorphic Lineagesand Striking BiogeographicPatterns

Similar to most major biological groups includ-ing birds (McKay et al 2013) fish (Wagner et al2013) plants (Griffin and Hoffmann 2014)arthropods (Schlick-Steiner et al 2010 Moreau2009) and many others finding and applying theappropriate character sets and analytical tools isone of the greatest challenges with empiricalspecies delimitation in lichen-forming fungi(Lumbsch and Leavitt 2011) Understanding thedifferences between morphological variationwithin a species and among closely related groupsis central to identifying diagnostic charactersrequired for establishing accurate phenotype-based taxonomic boundaries However in prac-tice a clear demarcation between intraspecificand interspecific variation is commonly subject toobservational bias and individual interpretation

Traditionally differences in morphologicalchemical and ecological features have been thepredominant source of diagnostic taxonomiccharacters for circumscribing lichen-formingfungal species (Printzen 2009) However lichensgenerally display few taxonomically useful char-acters relative to other groups (eg vascularplants vertebrates and arthropods) (Printzen2009) and varying levels of intraspecific variationamong different species groups may confoundaccurate taxonomic circumscriptions While somespecies may have little variation high levels ofintraspecific phenotypic variation are well docu-mented in some lichen-forming fungi (eg Xant-hoparmelia Hale 1990) Therefore molecular

2 The Dynamic Discipline of Species Delimitation hellip 13

genetic data now play a prominent role in delim-iting fungal species and understanding evolu-tionary relationships in lichens (Lumbsch andLeavitt 2011 Printzen 2009)

Arguably the use of molecular data has led toan improved perspective on the taxonomic valueof many phenotypic characters in lichenized fungiand species delimitation in general Cryptic spe-cies-level lineages are commonly identified usingmolecular data and in some cases these previ-ously unrecognized lineages are corroborated byformerly overlooked phenotypic characters (Pino-Bodas et al 2012a Spribille et al 2011 Divakaret al 2010 Arguumlello et al 2007) Other studieshave revealed the fact that some species-levellineages are likely comprised of chemically andmorphologically polymorphic individuals whichare conventionally considered as separate species(eg Leavitt et al 2011a Pino-Bodas et al 2011Velmala et al 2009) While these studies high-light the limitations of using traditional taxo-nomic characters for distinguishing naturalgroups in lichen-forming fungi they also providea valuable perspective on the importance ofongoing research in even the best-studied lichengroups Furthermore an improved perspective ofspecies boundaries has led to a strikingimprovement in understanding diversification anddistribution in many groups of lichenized fungi

Traditional phenotype-based approaches tospecies recognition appear to vastly underesti-mate diversity in lichen-forming fungi Whilemolecular research has corroborated traditionalphenotype-based hypotheses of species bound-aries in a number of cases studies repeatedlydemonstrate that our current interpretation ofmorphological and chemical characters oftenfails to accurately characterize species-leveldiversity (reviewed in Lumbsch and Leavitt2011) A growing body of evidence reveals that asignificant proportion of unknown diversityestimated for fungi including lichen-formingfungi is hidden under names of supposedlycommon and widespread species For exampleapproximately 80 unrecognized species-levellineages are estimated to occur in Parmeliaceae(Crespo and Lumbsch 2010) Even higher levels

of unrecognized species diversity are estimatedto occur in other families such as Graphidaceae(Rivas Plata and Luumlcking 2013 Rivas Plata andLumbsch 2011 Luumlcking 2012)

The topic of cryptic species cases wheretwo or more distinct species-level lineages areerroneously classified (and hidden) under onenominal taxon (Bickford et al 2007) has beenfrequently reviewed for lichen-forming fungi(Lumbsch and Leavitt 2011 Crespo and Lum-bsch 2010 Crespo and Peacuterez-Ortega 2009Printzen 2009) Although novel diagnostic phe-notypic characters may potentially be identifiedcorroborating ldquocrypticrdquo species these previouslyunrecognized lineages generally remain difficultto classify within a traditional phenotype-basedtaxonomy (Leavitt et al 2013c d) In somecases it appears that similar phenotypic charac-ters may arise in parallel at local or regionalscales but may not be correlated with naturalgroups or genetic isolation (Muggia et al 2014Rivas Plata and Lumbsch 2011 Grube andHawksworth 2007) For example in the cosmo-politan species Tephromela atra (Fig 21) up to15 independent lineages were identified usingphylogenetic analyses of molecular sequencedata However the continuum of morphologicaland chemical variability in the T atra complexcurrently prevents the description of new speciesusing traditional phenotype-based characters(Muggia et al 2014)

Recent research on the genus Cladonia(Cladoniaceae) highlights a fitting example of thecomplexities associated within using phenotypiccharacters for delimiting species in lichen-forming fungi (Fig 21 Pino-Bodas et al 20112012a b 2013a b) In the Cladonia gracilisgroup most currently accepted species were notrecovered as monophyletic clades and traditionaldiagnostic morphological characters were shownto be highly homoplasious (Pino-Bodas et al2011) Similarly C iberica and C subturgidahave been shown to constitute a single morpho-logically and chemically polymorphic species(Pino-Bodas et al 2012b) Similar patterns ofhigh degrees of morphological and chemicalpolymorphisms have also been observed in the

14 SD Leavitt et al

C cariosa group (Pino-Bodas et al 2012a) andC humilis species complex (Pino-Bodas et al2013a) High levels of intraspecific morphologi-cal and chemical polymorphisms are not restric-ted to Cladonia A clear demarcation betweenintraspecific and interspecific morphological andor chemical variation does not exist for a largeproportion of Xanthoparmelia species in westernNorth America and high intraspecific morpho-logical and chemical variation are common for anumber of species-level genetic groups (Fig 21

Leavitt et al 2011b c 2013e) In some cases asmany as eight traditionally circumscribed Xant-hoparmelia species were recovered in a singlespecies-level genetic group (Leavitt et al 2011c)High levels of intraspecific phenotypic variationhave been observed in a number of other generain Parmeliaceae including Bryoria (Velmalaet al 2009) Cetraria (Peacuterez-Ortega et al 2012)Vulpicida (Mark et al 2012) and others

The importance of biogeography in lichen-forming fungal evolution has remained somewhat

Fig 21 Examples of common lichens in which tradi-tional morphology-based species circumscriptions fail toreflect natural species-level fungal lineages a Cladoniagracilis photographed in the Clearwater Valley BritishColumbia Canada (see Pino-Bodas et al 2011) (photo-graph creditCurtis Bjoumlrk) bR shushanii a member of theRhizoplaca melanophthalma species group from the

Aquarius Plateau Utah USA (see Leavitt et al 2011a)(photograph credit S Leavitt) c Xanthoparmelia affwyomingica occurring in Coloradorsquos Front Range USA(see Leavitt et al 2011b c Leavitt et al 2013e) (photo-graph credit S Leavitt) d Tephromela atra sensu latofound in the Santa Monica Mountains California USA(see Muggia et al 2014) (photograph credit J Hollinger)


ambiguous due to the occurrence of phenotypi-cally similar lichens occurring across broadintercontinental distributions and uncertainty ofappropriate species circumscriptions While anumber of lichen-forming fungal species havebeen found to be truly widespread (eg Lindblomand Soslashchting 2008 Fernaacutendez-Mendoza et al2011 Ahti and Hawksworth 2005 Del-Pradoet al 2013) improved recognition of speciesboundaries has provided insight into importantbiogeographical patterns in lichens previouslyassumed to have cosmopolitan distribution pat-terns (Leavitt et al 2013d Del-Prado et al 2013Amo de Paz et al 2012 Seacuterusiaux et al 2011Otaacutelora et al 2010 Elix et al 2009) Crypticdiversity and complex biogeographic patterns arehighlighted in the Rhizoplaca melanophthalmaspecies group (Fig 21 Leavitt et al 2013d)Analyses of R melanophthalma sensu lato col-lected from five continents supported the presenceof cryptic species within this complex Two ofthese lineages were found to have broad inter-continental distributions while the other four werelimited to western North America (Leavitt et al2013d) Most strikingly of the six lineages fivewere found on a single mountain in the westernUSA and the sixth occurred no more than 200 kmaway from this mountain A recent study ofPseudocyphellaria (Lobariaceae) sensu lato inHawaii revealed a surprising number of previouslyunrecognized species hidden within nominal taxawith putative broad geographic distributionssuggesting a high degree of endemism in Hawaii(Moncada et al 2014) These studies providecrucial impetus to reevaluate species boundaries inorder to improve our understanding of diversitydistributions and evolution in lichenized fungi

22 A Practical Guideto Contemporary SpeciesDelimitation

As it has become clear that conventional phe-notype-based criteria for species circumscriptionsare often problematic (Bickford et al 2007)molecular data are thereby particularly valuablefor assessing traditional species boundaries and

for species delimitation in general (Lumbsch andLeavitt 2011 Fujita et al 2012) Below weprovide an overview of relevant operationalspecies delimitation methods the majority ofwhich are reliant on molecular data coupled withmolecular phylogenetic and population geneticanalyses Assuming that species do in fact rep-resent ldquosegments of metapopulation lineagesrdquo (deQueiroz 1998) direct genetic evidence of lineagestatus is particularly relevant to species delimi-tation studies when analyzed within a rigorousstatistical framework regardless of whether lin-eages differ in phenotypic characters that areapparent to human observers (Fujita et al 2012)This perspective should not be taken as supportfor disregarding phenotypic characters in speciesdelimitation studies (Edwards and Knowles2014 Fujita et al 2012 Yeates et al 2011)Rather hypotheses of species boundaries shouldbe considered more robust with increasing cor-roboration from independent data sources (iemolecular chemistry morphology ecologyetc) and the integration of independent data forempirical species delimitation studies should be amajor focus of species delimitation research(Fujita et al 2012)

Themajority of molecular phylogenetic studiesof lichenized fungi focus on generating sequencedata from a number of specimens representing thefocal group inferring a gene tree from the geneticdata matrix and assessing the monophyly of thesampled taxa This approach has provided valu-able acumen into evolutionary relationships thevalue of morphological characters for taxonomyand insight into diversity of lichenized fungi(Thell et al 2009 Reese Naeligsborg et al 2007Westberg et al 2007 Martiacuten et al 2003 Arup andGrube 2000 Lohtander et al 2000 Stenroos andDePriest 1998) However simply assessing themonophyly of traditional phenotype-based spe-cies often offers an incomplete perspective onspecies boundaries Studies explicitly designed forempirically delimiting species are pivotal toadvancing our understanding of speciation andspecies diversity in lichens For example althougha nominal species may be recovered as mono-phyletic in a gene tree intraspecific phylogeneticsubstructure may correspond to evolutionarily

16 SD Leavitt et al

independent lineages (eg Leavitt et al 2011a)As a construct of taxonomy recognizing amonophyletic clade comprised of multiple mor-phological indistinguishable species-level lin-eages as a single species may hold some appealhowever failing to formally recognize this diver-sity can have far-reaching implications in ourbiological (eg ecology evolution reproduction)interpretation of the group Alternatively well-supported intraspecific phylogenetic substructuremay be the result of stochastic evolutionary pro-cesses uniparental inheritance etc rather thanevolutionary independence Interpreting this typeof phylogenetic substructure as species-leveldiversity can introduce detrimental bias

Empirical species delimitation methods havebroadly focused on four main areas detectingputative species individual specimen assignmentto a species group or operational taxonomic unit(OTU) validation of candidate species or OTUsas evolutionarily distinct lineages and inferringspecies relationships (ie species tree inference)Ideally operational species delimitation criteriashould be based on explicit statistical protocolsin order to objectively assess species boundariesand minimize the need of subjective interpreta-tions or taxonomic expertise In this section weprovide a brief overview on a number of empir-ical species delimitation methods that to varyingdegrees fit these criteria For convenience wedivide these methods into three general catego-ries (i) species discovery methods for assigningsamples to populations without a priori infor-mation (ii) species validation approaches andcoestimating individual assignments and speciestrees and (iii) species delimitation using geno-mic data (Table 21)

Empirical species delimitation has receivedincreasing attention including ongoing develop-ment of bioinformatical approaches and themethods and programs provided in this chapterare by no means intended to be a comprehensivelist of all available analytical approaches Ratherthe methods provided here have been shown tobe useful in a number of previous studies orshow particular promise for future speciesdelimitation studies Our aim is that these can

serve as a starting point when designing studiesto assess species limits Not surprisingly variousapproaches to species delimitation may yielddifferent estimates of species boundaries and theresearcher may be required to make some degreeof subjective interpretation of the most biologi-cally appropriate species boundaries Comple-mentarily recently developed metrics forquantifying the congruence between two taxo-nomies (Ctax) and the potential for an approach tocapture a high number of species boundaries(Rtax) provide a means to objectively assess dis-crepancies among species delimitation methods(Miralles and Vences 2013) More sophisticatedapproaches including selection of speciesdelimitation models using approximate Bayesiancomputing (Camargo et al 2012 Fan andKubatko 2011 Beaumont et al 2010) anddesigning and conducting a simulation study thatmatches the characteristics of the empirical study(Carstens et al 2013 Camargo et al 2012) canbe used to more objectively evaluate competinghypotheses of species boundaries In most con-texts it is likely better to fail to delimit speciesthan it is to falsely circumscribe entities that donot represent actual species and therefore theinferences drawn from species delimitation stud-ies should generally be conservative (Carstenset al 2013 Miralles and Vences 2013)

221 Corroborating TraditionalTaxonomy and DiscoveringCryptic Species Using Single-Locus Data

Objectively defining a threshold separatingintraspecific population substructure from inter-specific divergence is the general aim of speciesdelimitation studies using single-marker datasetsMost species delimitation methods based onsingle-locus sequence data fall under two generalcategories either genetic distance or tree-basedapproaches (Sites and Marshall 2004) Distance-based approaches attempt to detect a differencebetween intra- and interspecific distances (ieldquobarcode gaprdquo) where the pairwise genetic


Table

21

Somemetho

dsused

forspeciesdelim

itatio

ninclud

ingspeciesdiscov

erymetho

dsforassign

ingsamples

topo

pulatio

nswith

outapriori

inform

ation(BAPS

Gaussianclustering

Guillo

trsquosUnified

Mod

elS

TRUCTURES

TRUCTURAMA)genetic

distance-based

metho

dforsortingsequ

encesinto

hypo

theticalspecies(A

BGD)tree-

basedspeciesdiscov

erymetho

ds(bGMYCG

MYCb

PTP

PTP

ldquoSpecies

Delim

itatio

nrdquoplug

-inforGeneiou

s)and

jointdiscov

eryandvalid

ationmetho

ds(BPamp

PBrownie

DISSE

CTSp

eDeStem)

Metho

dDescriptio

nInpu

tdata

BAPS

mdashpo

pulatio

nassignmentusing

Bayesianclustering

Aprogram

forBayesianinferenceof

thegenetic

structurein

apopulatio

nclustering

molecular

data

andperformingadmixture

analysesGenetic

mixture

analyses

canbe

performed

atbo

thgroupandindividu

allevelsusingeither

ano

n-spatialo

rspatialm

odel

BAPS

treatsboth

theallele

frequenciesof

themolecular

markers

(ornucleotid

efrequenciesforDNA

sequ

ence

data)andthenu

mberof

genetically

diverged

groups

inpopulatio

nas

random

variablesIn

theldquoclusteringwith

linkedlocirdquomod

elagenetic

mixture

analysiscanbe

done

usinghaploidsequ

ence

data

orotherlin

kedgenetic

markersA

nalysesandmodel

comparisons

canalso

beperformed

usingafixednu

mber

ofgenetically

diverged

grou

psor

prespecified

populatio

nstructures

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Availablefrom

httpwwwhelsinkifi

bsgsoftwareBAPS

describedin

Coran-

deret

al(200

420

0620

08)CoranderandMarttinen20

06)

Genotyp

icdatahaploidsequence

dataor

linkedmarkers

(AFL

Por

SNPs)

Gaussianclusteringmdashpo

pulatio

nassignment

usingGaussianclustering

Aprogram


thegenetic

structurein

apopulatio

nModelgroups

sampleinto

popu

latio

nsusinggeno

typicdata

bysearchingforclusters

that

canbe

attributed

tobeingmixturesof

norm

alallelefrequencydistributio

nsG

aussianclustering

requires

adatasetwhere

thecasesaredefinedby

variable

ofmetricscaleandhasbeen

used

with

geneticenvironmentalandmorphologicaldatasetsindividuallyinadditio

nto

integrated

datasets

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Implem

entedin

Rusingtheprabclus

(HausdorfandHennig20

10)andmclust

packages

(FraleyandRaftery

2007)

Genotyp

icdata

(flexible)

(con

tinued)

18 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Guillo

trsquosUnified

Mod

elmdashpo

pulatio

nassignmentusingBayesianclustering

Thisapproach

provides

astatistical

model

that

allowsoneto

analyzegenetic

and

phenotyp

ticdata

with

inaun

ified

mod

elandinferencefram

eworkandop

tionally

incorporateinform

ationaboutthespatialdistributio

nof

samplesA

Bayesianclustering

algorithm

assumes

that

each

clusterin

ageographical

domaincanbe

approxim

ated

bypo

lygons

that

arecentered

around

pointsgeneratedby

aPo

issonprocessGuillo

trsquos

Unified

Mod

elisflexiblein

term

sof

thegenetic

data

that

itcanutilize

andcapableof

accurately

delim

iting

species

Limita

tions

Genetic

andphenotypic

data

cantracedifferentevolutionary

historiesfor

instancephylogenetic

divergence

forneutralgenetic

markers

andadaptatio

nfora

morphological

structure

Source

Availableas

anextensionof

theRGENELAND

package(G

uillo

tet

al20

05

2012)(http

www2im

mdtudk

gigu

Geneland)

Genotyp

icandno

n-genetic

(eg

phenotypicalecolog

icalgeograph

ical

behavioral)data

STRUCTUREmdash

populatio

nassignment

usingBayesianclustering

Amodel-based

clustering

methodusingmultilocus

genotype

data

toinferpopulatio

nstructureandassign

individualsto

populatio

nsIndividualsin

thesampleareassigned

probabilistically

topopulatio

nso

rjointly

totwoor

morepopulatio

nsiftheirgenotypes

indicatethatthey

areadmixedT

hemodeldoes

notassum

eaparticular

mutationprocess

anditcanbe

appliedto

mosto

fthecommonly

used

genetic

markersp

rovidedthat

they

areno

tcloselylin

ked

The

methodcanproducehigh

lyaccurate

assignmentsusing

modestn

umbersof

loci(Pritchard

etal2

000)T

hemostappropriatelevelo

fpopulatio

nstructurecanbe

inferred

byassessinglik

elihoodscores

orthead

hocΔKstatistic

(Evann

oet

al20

05)

Limita

tions

Identifying

themostappropriatenumberof

genetic

clusters

ischallenging

clusters

produced

byST

RUCTUREcanbe

strongly

influenced

byvariationin

sample

sizeclusters

createdby

STRUCTUREmay

notbe

consistent

with

theevolutionary

historyof

thepopulatio

nswhentherearerelativ

elylong

divergence

times

with

inevolutionary

lineagesTem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnotexplicitlyestim

atedEquivalence

togenetic

clusters

tospecies-levelgroups

isuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httppritchardlabstanfordedustructurehtm

ldescribedin

Falush

etal(200

3)andPritchard

etal(200

0)

Genotyp

icdata

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim

izingHardyndashWeinb

ergdisequ

ilibrium

foragiven

partitioninglevelHow

everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition

inglevelNearlyanytype

ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram

aindexhtmldescribedin

Huelsenbeck

etal(201

1)

Genotyp

icdata

ABGDmdashbarcodegapusinggenetic

distances

ldquoAutom

aticBarcode

Gap

Discoveryrdquosortssequencesinto

hypotheticalspeciesbasedon

thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto

candidatespeciesthat

should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model

ofintraspecificbranchingwith

aYulemod

elforinterspecificbranching

which

isthen

fitto

aninferred

sing

le-gene

topology

toestim

atespeciesboundaries

andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san

inputtheGMYCapproach

requires

anultram

etricgene

tree

andrecent

refinementscanaccountforuncertaintyin

phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum

stancesincludingvariousmethods

ofphylogeneticreconstructio

nthepresence

ofahigh

numbersingletonshigh

numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim

itwell-supportedclades

ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto

rgprojectssplitshttpssitesgooglecom

site

noahmreidhom

esoftware)describedin

Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries

onagivenphylogenetic

inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof

substitutionsthereby

circum

ventingthepotentially

error-proneandcompute-intensive

processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F

urthermoreitappearsthatthePT

Pmodel

may

outperform

theGMYCandotherOTU-picking

methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpscoh-itsorgexelix

iswebsoftwarePT

Pindexhtmldescribedin

Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into

theGeneioussoftwareprovides

anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu

tesstatisticsrelatin

gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith

in-andbetween-speciesgenetic

distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas

user-selectedcollections

oftaxabutitdoes

not

providedefinitiv

esupportforspeciesgroupsItassumes

speciesaremonophyletic

Source

Implem

entedas

aplug-into

Geneious(geneiouscom)describedin

Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses

aBayesianmodelingapproach

togeneratethe

posteriorprobabilitiesof

speciesassignmentstaking

accountof

uncertaintiesdueto

unknow

ngene

treesandtheancestralcoalescent

processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio

nprobabilitiesitassumes

norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with

next-generationdatasetswith

100s

ofloci

Source

httpabacusgeneuclacuksoftwarehtmldescribedin

YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof

individualsandspeciestree

topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform

ofapriorthatisamodificatio

nof

thebirthndashdeathprior

forthespeciestreeItisim

plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently

describedmethodlackingathorough

theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce

httpcodegooglecompbeast-m

cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD

eSTEMmdashmultispecies

coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies

taxonomy(CarstensandDew

ey20

10)andcomputestheprobability

ofthegene

treesgiventhespeciestree

forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions

Accuracyisdependenton

quality

ofthegene

tree

estim

ates

Source

(http

carstenslaborgohio-stateedusoftwarehtml)criteria

describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom

differentlociassumingthatforaspeciatio

neven

thecorrespondingnodeson

gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree

andspeciesboundaries

remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic

programmingalgorithm

forestim

atingspeciestreesthat

bypasses

thecompu

tatio

nally

intensiveMCMCintegrationov

ergene

treesto

prov

idearigorous

techniqueforspeciesdelim

itatio

nstudiesusinggenome-wideSN

PdataCom

petin

gspeciesdelim

itatio

nmodelsarecomparedusingBayes

factorsanditappearsthat

this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce

httpwwwbeast2orgwikiindexphpBFD

describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata


distances among organisms belonging to thesame species are smaller than distances amongorganisms from different species (Puillandreet al 2012 Hebert et al 2003) Genetic distanceapproaches hold particular promise as an identi-fication tool shortcutting the difficulties of mor-phology-based identification (Hebert et al 2004)although in practice a barcode gap may not existfor many groups (Wiemers and Fiedler 2007)Furthermore the role of distance-based approa-ches using a single genetic marker for speciesdiscovery remains more controversial (Rubinoff2006) and without other corroborating evidenceOTUs inferred from single-locus dataset shouldonly be considered ldquocandidaterdquo species Tree-based methods aim to detect monophyletic cladescorresponding to species-level diversity bydetecting discontinuities associated within inter-and intraspecific branching patterns (Fujisawaand Barraclough 2013 Zhang et al 2013Monaghan et al 2009 Pons et al 2006) Tree-based species delimitation methods can also beused on the basis of other properties related tophylogenetic tree topologies (monophyly con-cordance with geography etc (Sites and Marshall2003 2004) Both distance- and tree-basedmethods have been applied for assessing speciesboundaries in lichen-forming fungi (Leavitt et al2012c 2014 Parnmen et al 2012 Del-Pradoet al 2010 2011 Molina et al 2011)

A number of tree-based methods partiallyautomate the species delimitation process withspecific bioinformatical analyses (Table 21)including the general mixed Yule coalescent(GMYC) approach (Fujisawa and Barraclough2013 Monaghan et al 2009 Pons et al 2006)and the Poisson tree processes (PTP) model(Zhang et al 2013) These methods provide rel-atively straightforward and objective pipelinesfor delimiting putative species-level lineagesfrom inferred gene trees by fitting within- andbetween-species branching models to an inferredsingle-locus topology A number of other tree-based methods for species delimitation areeffectively summarized in Sites and Marshall(2004)

When only single-locus data are available theGMYC has been shown to be a relatively robust

tool for species delimitation The GMYCapproach combines a coalescent model of intra-specific branching with a Yule model for inter-specific branching which is then fit to an inferredsingle-gene topology to estimate species bound-aries and a statistical measure of confidence for theinferred boundaries As an input the GMYCapproach requires an ultrametric gene tree andrecent refinements can account for uncertainty inphylogenetic relationships and parameters of theGMYC model (Fujisawa and Barraclough 2013Reid and Carstens 2012) Regardless of theseimprovements it may be difficult to accuratelyinfer an adequate ultrametric tree for large data-sets Although it appears that the GMYC isgenerally stable across a wide range of circum-stances including various methods of phyloge-netic reconstruction the presence of a highnumber singletons high numbers of sampledspecies and gaps in intraspecific sampling theaccuracy of the GMYC is most significantlyaffected by the mean population size relative todivergence times between them (Fujisawa andBarraclough 2013 Talavera et al 2013) Fur-thermore research suggests that the so-calledsingle-threshold version of the GMYC methodlikely outperforms the multiple-thresholdapproach (Fujisawa and Barraclough 2013Monaghan et al 2009) However other studiessuggest that theGMYCmethodmay often providehigher estimates for the total number ofOTUs thanother molecular species delimitation methods(Hamilton et al 2014 Miralles and Vences 2013Talavera et al 2013) warranting a cautiousinterpretation of results from GMYC analyses

Compared to the GMYC approach therecently introduced PTP method for speciesdelimitation has been suggested to be moreaccurate for preliminary species delimitation(Zhang et al 2013) Relative to the GMYC PTPoffers a more straightforward implementationrequiring a simple phylogenetic tree rather thanan ultrametric chronogram However moreresearch is required to assess the general perfor-mance of PTP across wide range of empirical andsimulated species delimitation studies At thispoint data support the use of the GMYC andPTP methods as objective and reasonable starting

24 SD Leavitt et al

points for species delimitation using single-genetopologies

The ldquoSpecies Delimitationrdquo plug-in to theGeneious software provides statistical approachesfor exploring species boundaries in single-genetopologies (Masters et al 2011) Using a priorispecimen assignments to putative species theldquoSpecies Delimitationrdquo plug-in computes statis-tics relating to the probability of the observedmonophyly or exclusivity having occurred bychance in a coalescent process and assesses thewithin- and between-species genetic distances inorder to infer the probability with which membersof a putative species might be identified suc-cessfully with tree-based methods An importantcontribution of the ldquoSpecies Delimitationrdquo plug-in is that it provides an objective means for usersto assess putative species within an empiricalstatistic-based framework rather than qualitativeevaluations of what level of hierarchy constitutesa species in a phylogeny

In contrast to simple sequence similaritythresholds (OTU-picking) for delimiting putativeevolutionarily significant units the AutomaticBarcode Gap Discovery (ABGD) method is anautomated procedure that sorts sequences intohypothetical species based on the existence of abarcode gap observed whenever intraspecificgenetic distances are less than those amongorganisms from different species (Puillandreet al 2012) Ultimately ABGD is a fast simplemethod that can be used to group individualsrepresented in a single-locus sequence alignmentinto candidate species that should be comple-mented with other lines of evidence in an inte-grative taxonomic approach (Kekkonen andHebert 2014 Puillandre et al 2012)

In general the ABGD GMYC and PTP ana-lytical protocols using single-locus data arerepeatable and computationally relatively fastproviding a valuable starting point for a pre-liminary perspective into species boundaries inunderstudied groups that can be validated withsubsequent studies (Kekkonen and Hebert 2014)Similarly analyses of single-locus data can beused to corroborate traditional phenotype-basedspecies boundaries and identify candidate speciesthat have previously been hidden within nominal

species However single-locus species delimita-tionmethodsmaybeparticularly prone to failure inrecognizing significant proportions of species-level biodiversity due to the strict criterion forreciprocal monophyly or alternatively providespurious inflations of estimated species diversitybased on genetic differences that donot correspondto species-level lineages For example recentestimates suggest that the incidence of species-level gene tree paraphyly is approximately 20 (Ross 2014) suggesting that analyses of single-locus datasets would likely fail to accurately deli-mit species in one in every five cases In contrastrecent empirical studies suggest that in some casesthe GMYC provides a striking overestimate ofspecies diversity (Miralles and Vences 2013)

The internal transcribed spacer region (ITS)has played a central role in molecular phyloge-netic studies of lichenized fungi and has recentlybeen adopted as the official barcoding marker forfungi (Schoch et al 2012) In many cases the ITSis sufficiently variable to resolve species bound-aries for lichenized fungi (Schoch et al 2012Kelly et al 2011) although accurate specimenidentification using sequence-based identificationapproaches requires a thoroughly curated refer-ence database (Leavitt et al 2014 Orock et al2012 Kelly et al 2011) In spite of the generalutility of the ITS marker a number of issues maypotentially limit its effective use in speciesdelimitation studies including the potential forintragenomic variation of the nuclear ribosomaltandem repeat and difficulties in aligning ITSsequences from divergent taxa (Kiss 2012)Because the ITS has been formally adopted as theofficial barcoding marker for fungi we recom-mend that species delimitation studies attempt toinclude this region for consistency across studiesHowever we recognize that in some taxonomicgroups the inclusion of the ITS may be prob-lematic and therefore suggest a careful screeningof other markers to identify appropriate loci forresolving species-level relationships

Ultimately the success of any single-locusspecies delimitation method largely depends onthe evolutionary history of the species groupunder study and the variability of the selectedmarker (Fig 22) Species delimitation will be


more difficult in more recently diverged lineagesand in cases with some level of interspecific geneflow relative to older well-diverged lineages(eg Leavitt et al 2012c 2013e) In the end wereemphasize that single-locus data alone shouldonly be used to provide a preliminary perspectiveof species boundaries and not as the sole evi-dence in species circumscriptions

222 Sampling Across the GenomeMultilocus Sequence Dataand Genome-Wide Markers

Although single-locus methods can provide anefficient approach for preliminary large-scaleassessments of species diversity there are

significant limitations particularly in recentlydiverged species groups where retention ofancestral polymorphisms and incomplete lineagesorting will likely result in different neutral locihaving unique gene topologies that do not mirrorthe speciation process (Knowles and Carstens2007 Heled and Drummond 2010 Taylor et al2000) In contrast to single-locus and strictlyphenotype-based approaches for species delimi-tation analyses of genetic data collected fromindependent genomic regions can provide robusthypotheses of species boundaries with increasingconfidence (Satler et al 2013 Zhang et al 2011Yang and Rannala 2010) Sequences from mul-tiple independent loci provide an importantsource of data for species delimitation studiesincluding recently developed models that

Fig 22 A simplified diagram illustrating the process ofspeciation through time in a single gene history andresulting gene topologies sampled at two points in thespeciation history (modified from Leliaert et al 2014)aEach dot represents a distinct gene copy and each row onenon-overlapping generation with lines connecting genecopies to their ancestors in the previous generation (onerow below) dashed diagonal lines represent reproductivebarriers two hypothetical sampling intervals at differentpoints in the speciation history are shown Tx and Ty

species delimitation methods using single-locus data areeffective only when species are reciprocally monophyleticin the sampled gene tree genetic clustering and coalescent-based species delimitation methods can circumscribespecies when species may not be monophyletic in sampledgenetic loci b Gene topology representing sampledhaplotypes at time Tx (shown in panel a) hypotheticalspecies are reciprocally monophyletic c Gene topologyrepresenting sampled haplotypes at time Ty (shown in panela) hypothetical species are para- and polyphyletic

26 SD Leavitt et al

combine individual gene genealogies and speciesphylogenies via modeling the coalescent historyof markers (Yang and Rannala 2010 Edwards2009 Knowles and Carstens 2007) As aresponse to advances in sequencing technologiesbioinformatical approaches for multimarker spe-cies delimitation analyses continue to be devel-oped (Table 21 Camargo and Sites 2013)

Long-term reproductive isolation of candidatespecies can be assessed with multilocus sequencedata by evaluating genealogical concordance ofunlinked markers (Baum and Shaw 1995 Aviseand Ball 1990) Within species the mixingeffects of recombination cause unlinked loci tohave distinct genealogical histories but geneticdrift and long-term divergence leads to concor-dant genealogical histories at loci across thegenome Relationships of individuals belongingto distinct candidate species can be evaluatedusing gene genealogies (eg haplotype networksor single-gene topologies) to identify lineagesthat exhibit genealogical exclusivity acrossunlinked neutral loci (Hudson and Coyne 2002Dettman et al 2003b Avise and Ball 1990) Thepresence of the same clades in the majority ofsingle-locus genealogies is taken as evidence thatthe clades represent reproductively isolated lin-eages (Dettman et al 2003a Pringle et al 2005)In practice the criteria of reciprocal monophylyand genealogical concordance of unlinked lociprovide a conservative approach for assessingspecies boundaries due to the fact that a sub-stantial amount of time is required after the initialdivergence of species until ancestral polymor-phisms have fully sorted (Knowles and Carstens2007 Hudson and Coyne 2002) Consequentlygroups with recent divergence histories willlikely go undiscovered under a genealogicalconcordance criterion due to the fact that speciesboundaries likely are not reflected in genegenealogies For example it would take morethan 1 million years after speciation before spe-cies would be delimited if 15 loci were sampledin species with an effective population size (Ne)of 100000 assuming one generation a yearunder a strict reciprocal monophyly criterion(Knowles and Carstens 2007) The amount oftime required for a species to be recognized using

a criterion of reciprocal monophyly increasesproportionally with increasing population sizes(Hudson and Coyne 2002) In a number ofstudies of lichen-forming fungi a genealogicalconcordance criterion has been used to bothdelimit previously unrecognized species-levellineages and validate some conventional pheno-type-based species (Leavitt et al 2012a c2013b Molina et al 2011 Kroken and Taylor2001)

Although analyses of molecular sequence datarelying on reciprocal monophyly or fixed char-acter differences for species delimitation canserve as important criteria for identifying speciesthese criteria are not commonly met acrossmultiple loci particularly in recent speciationhistories (Fig 22) To accommodate theobserved conflict among genealogies from mul-tiple loci with the underlying speciation historythe recent merge of coalescent theory with phy-logenetics has driven a major paradigm shift inspecies delimitation and molecular systematics ingeneral (Fujita et al 2012 Edwards 2009) Themultispecies coalescent model can be applied togenealogical histories from multiple independentloci to assemble separate coalescent processesoccurring in populations into a species tree(Degnan and Rosenberg 2009 Rannala andYang 2003) Within this coalescent-basedframework multiple individuals can be assignedto a single speciespopulation and the speciationhistory of ancestral and descendant species-levellineages can be inferred as a ldquospecies treerdquo incontrast to estimating gene genealogies fromindividual samples (Degnan and Rosenberg2006 2009 Rannala and Yang 2003)

A number of multispecies coalescent-basedspecies delimitation methods have recently beenintroduced offering an exciting framework forempirically assessing species boundaries byselecting the best species tree model from a set ofalternative models representing differenthypotheses of species boundaries (Table 21)For example SpeDeSTEM (Ence and Carstens2011) finds the maximum likelihood values for aspecies tree assuming all putative species areseparate lineages and for alternative species treeswhere two or more species are collapsed into a


single lineage The best fitting model is thenselected using the Akaike information criterionunder the assumption of constant populationsizes and fixed gene topologies across the speciestree The species delimitation program BayesianPhylogenetic and Phylogeography (BPampPYangand Rannala 2010) accommodates gene treeuncertainty and variable population sizes andsamples from the Bayesian posterior distributionof species delimitation models using reversible-jump Markov chain Monte Carlo (rjMCMC)sampling BPampP requires a user-provided guidetree with resolution finer than the species leveland evaluates alternative modes derived from allpossible subtrees that are generated by collapsingor splitting nodes on the guide tree (Yang andRannala 2010 Rannala and Yang 2003) Theproportion of time spent on each model is pro-portional to the posterior probability of themodel ie ldquospeciation probabilitiesrdquo WhileBPampP ranks among the most popular coalescent-based species delimitation programs (Fujita et al2012) misspecifications of the guide tree andorthe prior distribution for population size (θ) canresult in strong support for models containing aninflated number of species (Leacheacute and Fujita2010 Zhang et al 2011) The nonparametricheuristic species delimitation approach imple-mented in the program Brownie (OrsquoMeara 2010)jointly sorts anonymous samples into species andinfers a species tree from input gene trees fromdifferent loci assuming that for a speciationevent the corresponding nodes on gene trees willbe more consistent with each than the diver-gences within species However finding both theoptimum species tree and species boundariesremains computationally challenging andBrownie has been shown to frequently yieldincorrect results (OrsquoMeara 2010)

Most recently available coalescent-basedspecies delimitation methods require individualassignments to a species or population a priori Ina number of scenarios correct assignments ofsamples to species may be difficult includingpresence of cryptic species incongruencebetween conventional species and moleculardata or simply the fact that accurate specimenidentification in complex groups with subtle or

difficult to discern diagnostic characters is anontrivial task (Leavitt et al 2013e) Statisticalmethods for assessing individual assignment andspecies detection prior to coalescent-based spe-cies delimitation and species tree reconstructionand species provide a more objective approachfor understanding species boundaries

Population assignment tests using a variety ofclustering algorithms using genetic data offer auseful approach for grouping individuals intoputative reproductively isolated groups (Table 21Hausdorf and Hennig 2010 Corander et al2004 2008 Falush et al 2003 Pritchard et al2000) particularly for species delimitation prob-lems that exist at the interface of traditionalpopulation genetic and phylogenetic analyses(Carstens et al 2013 Weisrock et al 2010Knowles and Carstens 2007)

The programs STRUCTURE (Falush et al2003 Pritchard et al 2000) and STRUCTURA-MA (Huelsenbeck et al 2011) cluster samplesinto populations by minimizing HardyndashWeinbergdisequilibrium for a given number of populationclusters using unlinked genetic markers In gen-eral unlinked markers are not available for mostgroups of lichen-forming fungi and a numberof studies have used information from multilocussequence data for STRUCTURE analyses(Leavitt et al 2011a b c 2013e Fernaacutendez-Mendoza and Printzen 2013 Fernaacutendez-Mendoza et al 2011) Varying approaches havebeen implemented to convert DNA sequencedata to allelic data for Bayesian clustering (seeOrsquoNeill et al 2013) STRUCTURE is expectedto perform well when there is sufficient inde-pendence across regions such that linkage dis-equilibrium within regions does not dominate thedata (STRUCTURE manual) but can also beeffective using multilocus sequence data andtreating all SNPs as independent loci regardlessof physical linkage within each locus (OrsquoNeillet al 2013 Falush et al 2003) A recent study ofthe lichen-forming genus Letharia showed thatBayesian clustering implemented in the programSTRUCTURE was generally able to recover thesame putative Letharia lineages circumscribedemploying a gene genealogical approach inKroken and Taylorrsquos iconic species delimitation

28 SD Leavitt et al

study (Altermann et al 2014 Kroken and Taylor2001) Altermann et al (2014) show that popu-lation assignments were largely consistent acrossa range of scenarios including extensiveamounts of missing data the exclusion of SNPsfrom variable markers and inferences based onSNPs from as few as three gene regions Incontrast to STRUCTURE and STRUCTURA-MA the program BAPS (Corander et al 2008)can explicitly infer genetic structure using hap-loid sequence data or other linked geneticmarkers in the ldquoclustering with linked locirdquomodel Another advantage of BAPS is that theprogram is much more computationally efficientthan STRUCTURE or STRUCTURAMA Sim-ulation studies indicate that both BAPS andSTRUCTURE perform well at low levels ofpopulation differentiation and when clusters arenot well differentiated (Latch et al 2006)

In practice inferring the most appropriatelevel of population structure using Bayesianclustering algorithms remains challenging (Latchet al 2006 Evanno et al 2005) BAPS andSTRUCTURAMA can infer both individualassignments and the most likely number ofgenetic clusters (Corander et al 2008 Huelsen-beck et al 2011) and the ad hoc ΔK statisticproposed by Evanno et al (2005) provides anobjective approach for identifying the uppermosthierarchical level of structure However carefulconsideration of other levels of populationstructure may ultimately reveal more biologicallymeaning results and researchers should examineindividual assignments across a range of geneticgroupings Some markers generated fromindependent genomic regions such as SNPs andfast-evolving microsatellites can be used to dis-tinguish fine-scale population structuring on thebasis of allele frequencies and species delimita-tion analyses based on these markers may implythe risk of severe taxonomic oversplitting Inthese cases validation approaches such asBPampP and SpeDeSTEM and corroborationamong different species delimitation approachescan provide perspective between intraspecificpopulation structure and species-level clusters

Another limitation of clustering approaches isthat they do not assess or take into account

evolutionarily divergences and relationshipsamong population clusters Coalescent-basedspecies tree methods (discussed below) provide ameans to assess the diversification histories ofpopulations inferred from clustering analysesTherefore a potential working protocol for aninformed species delimitation study that takesinto account population structure could consist offirst applying a genetic clustering analysis under apopulation genetics criterion (eg BAPSSTRUCTURE STRUCTURAMA) to identifygenetically distinct population clusters that can beconsidered ldquocandidate speciesrdquo From these can-didate species a species tree can be inferred forfocal group using coalescent-based species treereconstruction methods (eg BEAST) Subse-quently a coalescent-based validation methodcan be applied to assess whether the distinctpopulation clusters represent independent evolu-tionary lineages (eg BPampP) This multistepapproach would provide a consistent and standardcriterion for distinguishing between population-and species-level lineages (Camargo and Sites2013) and has been applied to a number of casesincluding lichen-forming fungi (Leavitt et al2011a b c 2013e Leacheacute and Fujita 2010)

High-throughput sequencing methods providethe means to effectively sample hundreds tothousands of loci from across a species genome fora large number of species and continue to revolu-tionize studies that can be performed even in non-model organisms Targeted high-throughputsequencing approaches such as anchored phy-logenomics transcriptome sequencing reduced-representation genomic library sequencing(restriction-site-associated DNA sequencingRAD-Seq and genotype-by-sequencing GBS)and high-throughput amplicon sequencing pro-vide important insight into species boundaries fora number groups although none of these approa-ches has yet been applied to studies of lichenizedfungi

For example restriction-site-associated(RAD-Seq) DNA can simultaneously detect andgenotype thousands of genome-wide SNPs byfocusing the sequencing effort on a reducedrepresentation of the entire genome (Baird et al2008) and has been successfully applied to


intraspecies (Lewis et al 2007 Miller et al 2007Emerson et al 2010) and interspecies studies(Rubin et al 2012 Eaton and Ree 2013 Wagneret al 2013) This approach has provided strikinginsight into the recent adaption radiation of LakeVictoria cichlids (Wagner et al 2013) Analternative approach targeting the sequencing ofspecific loci using next-generation sequencingplatforms provides an efficient means of gener-ating data for loci of known genomic locationorthology size and expected level of variation(OrsquoNeill et al 2013) Markers can be targetedusing well-established PCR techniques or newerhybridization techniques and subsequentlypooled for high-throughput sequencing usingparallel-tagged sequencing of multiple individualsamples within a single sequencing run (OrsquoNeillet al 2013) A recent study of North Americantiger salamanders (Ambystoma tigrinum)demonstrated the potential for amplicon-basedparallel-tagged sequencing to rapidly generatelarge-scale genomic data for species delimitationand species tree research (OrsquoNeill et al 2013)

Although these methods provide an enormousamount of data and insight into diversification atan unprecedented scale computational limita-tions restrict the applicable analytical methodsfor large datasets although computational andanalytical advances are happening rapidly Mostrecent coalescent-based species delimitation andspecies tree models using gene trees have beenlimited to handle tens of loci for multiple indi-viduals and combining hundreds or thousands ofgene trees into a single species delimitationframework presents considerable computationalchallenges (Camargo and Sites 2013) Therecently developed method Bayes factor delim-itation (with genomic data BFD) combines adynamic programming algorithm for estimatingspecies trees that bypasses the computationallyintensive MCMC integration over gene trees toprovide a rigorous technique for species delimi-tation studies using genome-wide SNP data(Leacheacute et al 2014) Competing species delimi-tation models are compared using Bayes factorsand it appears that this approach is robust tosample sizes (ie few loci and limited samples

per species) and misspecification of the prior forpopulation size (θ) (Leacheacute et al 2014)

Fungi are an ideal model for assessing diver-gence in eukaryotes due to their simple mor-phologies small genomes broad ecological rolesand diverse lifestyles (Gladieux et al 2014)However the use of genomic data from high-throughput sequencing methods have not yet beenincluded in species delimitation studies of lichen-forming fungi This is due in part to challenges indealing with metagenomic data containing geno-mic information from a plethora of symbiontsassociated with a lichen thallus scant genomicresources and discipline-specific inertia A num-ber of lichen-forming fungal genomes are cur-rently available including Endocarponpusillum(Wang et al 2014) Cladonia spp (Park et al2013a 2014) Caloplaca flavorubescens (Parket al 2013b) and Cladonia greyi and Xanthoriaparientina through the Fungal Genomics Programat Joint Genome Institute of the United StatesDepartment of Energy (httpwwwjgidoegov)The foreseeable ongoing expansion of genomicresources for lichen-forming fungi will be centralto developing approaches for delimiting speciesusing high-throughput sequencing Specificallygenomic resources will be crucial in identifyingnovel markers (variable genes regions SNPsmicrosatellites etc) identifying conserved fungalmarkers for targeted high-throughput sequencingapproaches and references for entire genomecomparisons (Devkota et al 2014 Werth et al2013)

23 Can We Make SpeciesDelimitation in Lichen-FormingFungi Truly Integrative

The widespread availability of genetic data hascreated a biased viewpoint that only genetic datashould be used for statistical species delimitationHowever an ongoing appeal to researchers toassess species boundaries from multiple andcomplementary perspectives (phylogeneticspopulation genetics comparative morphologydevelopment ecology etc) has resulted in an

30 SD Leavitt et al

integrative taxonomic framework bringing theseconceptual and methodological perspectivestogether (Hamilton et al 2014 Fujita et al 2012Salicini et al 2011 Yeates et al 2011 Padialet al 2010 Dayrat 2005 Will et al 2005 Wiensand Penkrot 2002) In practice any study linkingdifferent kinds of data to support hypotheses ofspecies boundaries including mapping morpho-logical characters onto a molecular phylogenycan be considered integrative Integrative meth-ods for species delimitation fall across a broadspectrum ranging from verbal and qualitativeassessments of data classes to quantitative meth-ods that allow different data types to contribute tostatistical species delimitation (Yeates et al2011) In most taxonomic studies utilizing bothmolecular and morphological data expert opinionis eventually used to some degree to evaluate thefinal status of a candidate species (eg Bond andStockman 2008) In this sense many studiesusing evidence for independent data sources fordelimiting species boundaries use an iterativeapproach (Yeates et al 2011) where speciesboundaries can be tested within a hypothetico-deductive framework with diverse datasets

From a practical perspective we advocate aprocess of iterative taxonomy (sensu Yeates et al2011) to circumscribe and refine species limitsusing multiple lines of evidence This iterativeprocess involves comparisons of morphologicaldata with a phylogenetic hypothesis (ie single-locus gene tree or concatenated multilocusphylogeny) to identify the least inclusive mono-phyletic clade in the topology characterized by atleast one unambiguously diagnostic morphologi-cal character (Miralles and Vences 2013) Thisphylogenetic tree-informed approach to assessingspecies boundaries represents a common practicein studies of lichen-forming fungi Previouslyunrecognized species-level clades with corre-sponding subtle or overlooked phenotypiccharacters have been commonly observed in bothcrustose lichens [eg Acarospraceae (Wedin et al2009) Graphidaceae (Rivas Plata and Lumbsch2011 Papong et al 2009) Lecanoraceae (Leavittet al 2011a) Lecideaceae (Ruprecht et al 2010)Mycoblastaceae (Spribille et al 2011) andTeloschistaceae (Vondraacutek et al 2009 Muggia

et al 2008)] and foliose and fruticose lichens [egLobariaceae (Moncada et al 2014 McDonaldet al 2003) Parmeliaceae (Leavitt et al 2013aWirtz et al 2012 Divakar et al 2005 Molinaet al 2004 Kroken and Taylor 2001) Peltigera-ceae (OrsquoBrien et al 2009 Goffinet et al 2003)Physciaceae (Elix et al 2009 Divakar et al2007) and Sphaerophoraceae (Houmlgnabba andWedin 2003)] As a specific example Luumlckinget al (2008) used a combination of medullarychemistry and underside pigmentation in speci-mens from the Heterodermia obscurata group inCosta Rica to corroborate monophyletic clades inan ITS gene tree as species-level lineages

While this iterative approach to speciesdelimitation and taxonomy has proven valuablefor understanding species boundaries anddescribing new taxa in some groups of lichenizedfungi a posteriori examination of morphologicaland chemical features has failed to reveal diag-nostic phenotypic characters in a number ofstudies (Muggia et al 2014 Leavitt et al 2011aOtaacutelora et al 2010 OrsquoBrien et al 2009) Fur-thermore a study of widespread species in thegenus Melanelixia (Parmeliaceae) indicated thatphenotypically cryptic lichen-forming fungalspecies-level lineages may be relatively ancientand diagnosable phenotypic differences may beabsent even millions of years after the initialdivergence (Leavitt et al 2012b) The latter studyhighlights the fact that species-level lineages maycommonly exist without any observable diag-nostic phenotypic characters calling into ques-tion the impetus for a universal application ofintegrative taxonomy

In keeping with the principle that as manylines of evidence as available should be com-bined to delimit species (Dayrat 2005) the for-malized integrative taxonomic approach (ITAXMiralles and Vences 2013) uses a mtDNA guidetree and observations from different types of datathat might provide conclusive evidence for theindependence of lineages and thus their identityas different species Miralles and Vences (2013)provide a non-exhaustive list of species delimi-tation criteria to be integrated in the ITAXapproach including (a) sympatric occurrencewithout admixture as revealed by consistent


differences in morphological or molecular char-acters at the same geographic location (b) strongdifferences in a behavioral morphological orgenetic character known to mediate prematingisolation (c) unviability or infertility of hybrids(d) lack of gene flow across a geographicalhybrid zone (e) congruent diagnostic differencesbetween sister lineages in various unlinkedmorphological character (f) absence of haplo-type sharing in several unlinked nuclear loci and(g) a combination of criteria endashf Speciesboundaries are based on seeking the least inclu-sive monophyletic group in the mtDNA treewhich fulfills at least one of the criteria listedabove Clearly the ITAX approach is sensitive tosample size and in order to reliably support thedistinctiveness of a given species it has beenrecommended that the sampling strategy includesat least five individuals per species (Miralles andVences 2013) A similar approach could beadopted for lichenized fungi using an ITS genetopology as guide tree rather than mtDNA

A total evidence approach including concat-enation has been a common approach for inte-grating information from different sourcesincluding independent genetic markers in phy-logenetic reconstructions (Kluge 1989 Wiens1998 de Queiroz et al 1995) Proponents forconcatenation of independent data in phyloge-netic analyses argue that when combining alldata the underlying signal of speciation mayemerge even if weakly contradictory signalis contained in the individual data partitions(Gatesy et al 1999) For example establishing apreliminary perspective of species boundariesfrom multilocus sequence data using concatena-tion may provide a reasonable starting point forscreening for cryptic species and species treeinference (Šlapeta et al 2006 Leavitt et al2011a Le Gac et al 2007 Leacheacute 2009)However concatenation and consensus methodsimply a risk of obtaining inflated support forincorrect relationships and information aboutvariance in gene coalescence is lost (Degnan andRosenberg 2006 2009 Kubatko and Degnan2007 Edwards 2009) In spite of the impetus forintegrating evidence from independent data

sources into empirical species delimitation stud-ies and taxonomy (Fujita et al 2012 Padial et al2009) most currently available species delimi-tation methods are unable to accommodate non-genetic data sources in a statistical framework

231 Selecting the Appropriate Data

In the face of increasing availability of geneticdata and associated bioinformatical approachesfor delimiting species researchers should care-fully consider what information is being sacri-ficed by the failure to consider non-genetic datain species delimitation studies and whetheraccuracy could be improved by the addition ofmultiple data types Morphological data havehistorically served as a proxy to identify repro-ductively isolated groups (ie ldquospeciesrdquo) (Ray1686 Fujita et al 2012) Current methods fordelimiting species using non-genetic data (egchemistry morphology and ecology) remainwoefully understudied For example morphol-ogy-based species circumscriptions are generallybased on one or more qualitative (or quantitative)morphological characters that do not appear tooverlap with other species However ascertain-ing that a given trait is truly fixed within apopulation with statistical confidence requiresunrealistic sample sizes even when allowing forsome level of polymorphism in the diagnosticcharacter (Wiens and Servedio 2000) Now var-ious combinations of datamdashfrom morphologygenetics geography and ecologymdashare acceptedas standard information for species delimitationstudies (Ruiz-Sanchez and Sosa 2010 Ross et al2010 Edwards and Knowles 2014)

Below we briefly discuss appropriate datasources for species delimitation studies of lichen-forming fungi However homoplastic characters(similar traits that are not derived from a com-mon ancestor) are common among many traitscommonly used to circumscribe fungal taxa thebiological significance of secondary metabolitevariation remains largely unknown (Lawrey1986) and ecological niches may be difficultto adequately characterize and model due to

32 SD Leavitt et al

microhabitat requirements and data resolutionTherefore we advocate a cautious approach toselecting appropriate and relevant data forassessing species boundaries

In spite of the potential challenges and limi-tations of using phenotypic data these traits haveprovided a plethora of valuable information forunderstanding species boundaries For examplein the Melanelixia fuliginosa group (Parmelia-ceae) a morphometric analysis using color isi-dia and marginal zone free of isidia ascharacters revealed a general pattern of differ-entiation between material formerly recognizedas subspecies fuliginosa and glabratula in spiteof the fact that considerable overlap betweengroups occurs in some characters individually(Arup and Berlin 2011) Furthermore the dis-tinction of the two groups was supported by aphylogenetic analysis of the ITS marker andecological differences with M fuliginosaoccurring predominantly on rock and M glab-ratula on bark (Arup and Berlin 2011) Thisstudy of the M fuliginosa group provides a fit-ting example of using multiple independentsuites of data including ecology morphologyand genetic information to establish a robusthypothesis of species boundaries

In lichen systematics phenotypic dataincluding thallus organization secondary metab-olites mode of reproduction ascoma-type andontogeny ascus and ascospore characters havehistorically played a prominent role (Printzen2009) Ascomatal characters have traditionallyheld a major role in higher-level classification(Printzen 2009) in contrast to species-levelclassification which also tends to include a widearray of vegetative and chemical charactersCommonly assessed morphological charactersinclude thallus form and size cortical features(eg maculae and pseudocyphellae) presenceformcolor of attachment structure (eg rhizines)and reproductive mode (ascomata vs vegetativediaspores) (Printzen 2009) Different morpho-logical types of vegetative diasporesmdashcorticatedisidia and ecorticated sorediamdashand their locationare commonly used to distinguish speciesAscomatal characters including morphologylocation (laminal vs marginal) position (sessile

vs immersed) presence of thalline margins colorof an apothecial disc and presence or color ofpruina are also commonly used in speciesdelimitations of lichen-forming fungi (Printzen2009) Other important characters may includethalline characters form color size and septationof ascospores size and form and structure of ascithe hamathecium type of epihymenium andhypothecium and the type of excipulum orperidium conidiomatal characters etc (Printzen2009)

Assessments of secondary metabolites hasplayed an important role in lichen taxonomybeginning with the introduction of simple spottests by Nylander (1866a b) The use of chem-istry in lichen taxonomy has been discussed indetail in numerous reviews (Lumbsch 1998a bRogers 1989 Brodo 1986 Egan 1986 Leuckert1985 Brodo 1978 Hawksworth 1976 Culber-son 1969 1970) and we refer readers to thesevaluable sources for a more comprehensive per-spective on lichen chemistry In short extrolites(secondary metabolites) belong to various clas-ses the most common and diverse include dep-sides depsidones chlorinated xanthones andanthraquinones (Lumbsch 2002 Culberson1969) The presence or absence of specificextrolites or their replacements by another sub-stance is widely used to distinguish speciesparticularly when correlated with differences ingeographic distributions However if morpho-logical or geographical differences betweenpopulations containing different extrolites are notapparent the taxonomic significance has beendisputed with some authors distinguishing themas species and others preferring to regard them aschemical races within a species In addition tosimply using the presence or absence of extro-lites Culberson and Culberson (1976) proposedto arrange lichen substances into chemosyn-dromes of closely related substances The pres-ence or absence of these chemosyndromes maypotentially be used as characters to delimit spe-cies regarding differences involving the samechemosyndromes as intraspecific variation anddistinct chemosyndromes as evidence for inter-specific populations (Lumbsch 1994 Gowan1986)


Due to the fact that species delimitationstudies often incorporate a substantial biogeo-graphical or ecological component ecologicalniche modeling plays an increasing role in phy-logenetic and taxonomic research Niche mod-eling can provide evidence for ecologicalisolation between populations based on eitherconserved or divergent ecological niches andtherefore can provide additional evidence sup-porting lineage independence between putativespecies The application of ecological nichemodeling has been applied to species delimita-tion studies in a number of cases (Ruiz-Sanchezand Sosa 2010 Leacheacute et al 2009 Raxworthyet al 2007 Rissler and Apodaca 2007) althoughit has not been explicitly applied to assess speciesboundaries in lichen-forming fungi By mappingthe spatial distribution of environmental suit-ability of climatic variables for candidate speciesthe application of ecological niche modeling canbe particularly important in cases where specieshave allopatric distributions (Raxworthy et al2007)

Ecological niche modeling utilizes knownassociations between a speciesrsquo occurrencelocalities and environmental variables to defineabiotic conditions within which populations canbe maintained (Guisan and Thuiller 2005) Themethodological approach for modeling is basedon four general properties (i) The current knownspeciesrsquo localities is the dependent variable (ii)the distribution is modeled as a map composed ofgrid cells at a specified resolution (iii) a range ofenvironmental variables (eg temperature pre-cipitation and solar exposure) are collected todescribe the characteristics of each cell and (iv)classifying the degree to which each cell is eithersuitable or unsuitable for each species under arange of models (Guisan and Thuiller 2005)

Ecological data also have the potential to playan important role in understanding speciesboundaries in lichen-forming fungi For exampleNash and Zavada (1977) demonstrated that Xant-hoparmelia populations with distinct chemistriesoccurring in the northern portion of the SonoranDesert exhibit habitat selection among differentrock substrates within a region with relativelyuniform climate and topography In another case

the parmelioid species Parmelia mayi is mor-phologically indistinguishable from P saxatilisbut can be separated by bioclimatic features andgenetic and chemical characters (Molina et al2011) McCune and Printzen(2011) assess distri-butions and climatic niches of species in theLecanora varia group inwesternUSAandprovidea model that uses continental influence and annualtemperature as themajor factors predicting speciesdistributions The distribution of Usnea hirta alichen commonly used in air quality biomonitor-ing research was modeled for a section of theWhite River National Forest in central Coloradobased on the presence of U hirta at 72 biomoni-toring reference sites distributed in the inter-mountain western United States (Shrestha et al2012) The best model for predicting U hirtadistribution included four variablesmdashsolar radia-tion average monthly precipitation and averagemonthly minimum and maximum temperatures(Shrestha et al 2012) These studies supportthe potential use of ecological niche modelingmethods in species delimitation studies of lichen-forming fungi

24 Conclusions What AboutTaxonomy

In most cases species circumscription and tax-onomy requires some degree of qualitativejudgment and individual interpretation Integrat-ing multiple types of data into an empiricalframework for delimiting species boundarieswhere species boundaries can be tested within ahypothetico-deductive framework with diversedatasets can provide robust hypotheses of speciesboundaries and taxonomic stability (Yeates et al2011) However it has long been known toevolutionary biologists that distinct species donot need to have diagnosable morphologicaldifferences (Mayr 1963) and increasing avail-ability of genetic data has allowed researchers toidentify species and to rigorously test speciesboundaries with a level of precision that wasunimaginable a decade ago While analyticaladvances in statistical species delimitation havebeen largely based on genetic data the utility of

34 SD Leavitt et al

these approaches to formal taxonomy remainselusive Due to these challenges an eclecticapproach to delimiting species and cautionagainst the reliance on any single dataset ormethod is required when delimiting species

While results from many recent studies maycontradict traditional species boundaries acrossmany groups of lichen-forming fungi we areoptimistic that this research represents substantialprogress toward a more accurate perspective onspecies boundaries and diversity in fungi As aresult of ongoing research of species boundariesin lichen-forming fungi the taxonomic value ofmany phenotypes is now better understood ourunderstanding of ecological evolutionary andbiogeographic patterns has improved and we canbegin to better understand patterns of symbiontinteractions in lichens Integrating new data(including novel morphological characters andgenetic data) will be essential to accurately rep-resent species-level diversity across all groups oflichenized fungi Hopefully an improved per-spective on lichen diversity also increases ourappreciation of these incredible symbiotic sys-tems While we are strong advocates for theapplication of independent data types in devel-oping an integrative taxonomy there is anincreasing need to formally recognize the exis-tence of phenotypically cryptic species-levellineages in lichen-forming fungi (see Hibbettet al 2011) In some cases a molecular taxon-omy may provide the most practical approach toconsistent treatment of mycobiont species withinlichen groups where diagnostic morphologicalcharacters are unidentifiable or practically notfeasible (Leavitt et al 2013c d)

These are exciting times for taxonomists andphylogeneticists A closer look at lichen taxon-omy with the inclusion of new data will help usto better understand the diversity of these fasci-nating organisms accurately interpret distribu-tion patterns and play a more important role inmeaningful conservation practices Howeversome level of uncertainty will accompany pro-gress In many taxonomic groups our traditionalapproach for species identification will likelyneed to be substantially modified The search forcorroborating morphological support for cryptic

species identified using molecular data willrequire meticulous and creative approaches toassess phenotypic variation in potentially unor-thodox ways We are hopeful that lichenologistswho traditionally have been eager to include newmethods such as chromatography in their rou-tine identifications will be amenable to includemolecular techniques to their routine examina-tion of specimens for identification and classifi-cation Although this may prove difficult toachieve by single individuals especially citizenscientists that traditionally play an important rolein lichen taxonomy the increasing numberof collaborative projects in the discipline (egLumbsch et al 2011 Crespo et al 2010Gueidan et al 2009) make us optimistic thatbroad-scale collaborative approaches will facili-tate the inclusion of molecular data in lichenresearch at all levels

Acknowledgments We are indebted to various col-leagues for valuable thought-provoking discussionnotably Matthew Nelsen (University of Chicago) AnaCrespo (Universidad Complutense de Madrid) PradeepDivakar (Universidad Complutense de Madrid) BeckettSterner (The Field Museum) and Joyce Havstad (TheField Museum) We also thank anonymous reviewers whoprovided valuable comments that improved this chapterSupport by the US National Science Foundation isgratefully acknowledged (ldquoHidden diversity in parmelioidlichensrdquo DEB-0949147)

References

Ahti T Hawksworth DL (2005) Xanthoparmelia steno-phylla the correct name for X somloeumlnsis one of themost widespread usnic acid containing species of thegenus The Lichenologist 37(4)363ndash366 doi101017S0024282905015197

Altermann S Leavitt SD Goward T Nelsen MPLumbsch HT (2014) How do you solve a problemlike Letharia A new look at cryptic species in lichen-forming fungi using Bayesian clustering and SNPSsfrom multilocus sequence data PLoS ONE 9(5)e97556 doi101371journalpone0097556

Amo de Paz G Cubas P Crespo A Elix JA Lumbsch HT(2012) Transoceanic dispersal and subsequent diver-sification on separate continents shaped diversity ofthe Xanthoparmelia pulla group (Ascomycota) PLoSONE 7(6)e39683 doi101371journalpone0039683

Arguumlello A Del Prado R Cubas P Crespo A (2007)Parmelina quercina (Parmeliaceae Lecanorales)


includes four phylogenetically supported morphospe-cies Biol J Linn Soc 91(3)455ndash467 doi101111j1095-8312200700810x

Arup U Berlin ES (2011) A taxonomic study ofMelanelixia fuliginosa in Europe The Lichenologist43(02)89ndash97 doi101017S0024282910000678

Arup U Grube M (2000) Is Rhizoplaca (Lecanoraleslichenized Ascomycota) a monophyletic genus Can JBot 78(3)318ndash327 doi101139b00-006

Avise J Ball R (1990) Principles of genealogicalconcordance in species concepts and biological tax-onomy Oxf Surv Evol Biol 745ndash67

Baird NA Etter PD Atwood TS Currey MC Shiver ALLewis ZA Selker EU Cresko WA Johnson EA(2008) Rapid SNP discovery and genetic mappingusing sequenced RAD markers PLoS ONE 3(10)e3376 doi101371journalpone0003376

Baum DA Shaw KL (1995) Genealogical perspectives onthe species problem In Hoch PC Stephenson AG(eds) Experimental and molecular approaches to plantbiosystematics Missouri Botanical Garden St Louispp 289ndash303

Beaumont MA Nielsen R Robert C Hey J Gaggiotti OKnowles L Estoup A Panchal M Corander JHickerson M Sisson SA Fagundes N Chikhi LBeerli P Vitalis R Cornuet J-M Huelsenbeck J FollM Yang Z Rousset F Balding D Excoffier L (2010)In defence of model-based inference in phylogeogra-phy Mol Ecol 19(3)436ndash446 doi101111j1365-294X200904515x

Bickford D Lohman DJ Sodhi NS Ng PKL Meier RWinker K Ingram KK Das I (2007) Cryptic speciesas a window on diversity and conservation TrendsEcol Evol 22(3)148ndash155 doi101016jtree200611004

Bonan GB Shugart HH (1989) Environmental factors andecological processes in boreal forests Annu Rev EcolSyst 20(1989)1ndash28

Bond J Stockman A (2008) An integrative method fordelimiting cohesion species finding the population-species interface in a group of Californian trapdoorspiders with extreme genetic divergence and geo-graphic structuring Syst Biol 57(4)628ndash646 doi10108010635150802302443

Brodo IM (1978) Changing concepts regarding chemicaldiversity in lichens The Lichenologist 10(1)1ndash11doi101017S0024282978000031

Brodo IM (1986) Interpreting chemical variation inlichens for systematic purposes The Bryologist 89(2)132ndash138

Caley MJ Fisher R Mengersen K (2014) Global speciesrichness estimates have not converged Trends EcolEvol 29(4)187ndash188 doi101016jtree201402002

Camargo A Sites JW (2013) Species delimitation adecade after the Renaissance In Pavlinov I (ed) Thespecies problemmdashongoing issues InTech doi10577252664

Camargo A Morando M Avila LJ Sites JW (2012)Species delimitation with ABC and other coalescent-based methods a test of accuracy with simulations and

an empirical example with lizards of the Liolaemusdarwinii complex (Squamata Liolaemidae) Evolution66(9)2834ndash2849 doi101111j1558-5646201201640x

Carstens BC Dewey TA (2010) Species delimitationusing a combined coalescent and information-theoreticapproach an example from North American Myotisbats Syst Biol 59(4)400ndash414 doi101093sysbiosyq024

Carstens BC Pelletier TA Reid NM Satler JD (2013)How to fail at species delimitation Mol Ecol 22(17)4369ndash4383 doi101111mec12413

Corander J Marttinen P (2006) Bayesian identification ofadmixture events using multi-locus molecular mark-ers Mol Ecol 15(10)2833ndash2843 doi101111j1365-294X200602994x

Corander J Waldmann P Marttinen P Sillanpaa M(2004) BAPS 2 enhanced possibilities for the analysisof genetic population structure Bioinformatics 20(15)2363ndash2369 doi101093bioinformaticsbth250

Corander J Marttinen P Mantyniemi S (2006) Bayesianidentification of stock mixtures from molecular markerdata Fish Bull 104550ndash558

Corander J Marttinen P Siren J Tang J (2008) EnhancedBayesian modelling in BAPS software for learninggenetic structures of populations BMC Bioinf 9(1)539 doi1011861471-2105-9-539

Coyne JA Orr HA (2004) Speciation Sinauer AssociatesSunderland

Cracraft J (1983) Species concepts and speciation anal-ysis Curr Ornithol 1159ndash187

Crespo A Lumbsch HT (2010) Cryptic species in lichen-forming fungi IMA Fungus 1167ndash170 doi105598imafungus2010010209

Crespo A Peacuterez-Ortega S (2009) Cryptic species andspecies pairs in lichens a discussion on the relation-ship between molecular phylogenies and morpholog-ical characters Anales del Jardin Botanico de Madrid66(S1)71ndash81 doi103989ajbm2225

Crespo A Kauff F Divakar PK del Prado R Perez-Ortega S Amo de Paz G Ferencova Z Blanco ORoca-Valiente B Nunez-Zapata J Cubas P ArguelloA Elix JA Esslinger TL Hawksworth DL MillanesA Molina MC Wedin M Ahti T Aptroot A et al(2010) Phylogenetic generic classification of parmeli-oid lichens (Parmeliaceae Ascomycota) based onmolecular morphological and chemical evidenceTaxon 59(6)1735ndash1753

Culberson WL (1969) The use of chemistry in thesystematics of the lichens Taxon 18498ndash505

Culberson WL (1970) Chemosystematics and ecology oflichen-forming fungi Annu Rev Ecol Syst 1153ndash170

Culberson CF Culberson WL (1976) Chemosyndromicvariation in lichens Syst Bot 1325ndash339

Darwin C (1859) On the origin of species by means ofnatural selection or the preservation of favoured racesin the struggle for life J Murray London

Dayrat B (2005) Towards integrative taxonomy Biol JLinn Soc 85(3)407ndash415 doi101111j1095-8312200500503x

36 SD Leavitt et al

de Queiroz K (1998) The general lineage concept ofspecies species criteria and the process of speciationa conceptual unification and terminological recom-mendations In Howard DJ Berlocher SH (eds)Endless forms species and speciation Oxford Uni-versity Press Oxford pp 57ndash75

de Queiroz K (1999) The general lineage concept ofspecies and the defining properties of the speciescategory In Wilson RA (ed) Species new interdis-ciplinary essays MIT Press Cambridge pp 49ndash89

de Queiroz K (2007) Species concepts and speciesdelimitation Syst Biol 56(6)879ndash886 doi10108010635150701701083

de Queiroz A Donoghue MJ Kim J (1995) Separateversus combined analysis of phylogenetic evidenceAnnu Rev Ecol Syst 26(1)657ndash681 doi101146annureves26110195003301

Degnan J Rosenberg N (2006) Discordance of speciestrees with their most likely gene trees PLoS Genet 2(5)e68 doi101371journalpgen0020068

Degnan JH Rosenberg NA (2009) Gene tree discordancephylogenetic inference and the multispecies coales-cent Trends Ecol Evol 24(6)332ndash340 doihttpdxdoiorg101016jtree200901009

Del-Prado R Cubas P Lumbsch HT Divakar PK BlancoO de Paz GA Molina MC Crespo A (2010) Geneticdistances within and among species in monophyleticlineages of Parmeliaceae (Ascomycota) as a tool fortaxon delimitation Mol Phylogenet Evol 56(1)125ndash133 doi101016jympev201004014

Del-Prado R Divakar PK Crespo A (2011) Using geneticdistances in addition to ITS molecular phylogeny toidentify potential species in the Parmotrema reticul-atum complex a case study The Lichenologist 43(06)569ndash583 doi101017S0024282911000582

Del-Prado R Blanco O Lumbsch HT Divakar PK ElixJA Molina MC Crespo A (2013) Molecular phylog-eny and historical biogeography of the lichen-formingfungal genus Flavoparmelia (Ascomycota Parmelia-ceae) Taxon 62(5)928ndash939 doi101270562522

Dettman J Jacobson D Taylor J (2003a) A multilocusgenealogical approach to phylogenetic species recog-nition in the model eukaryote Neurospora Evolution57(12)2703ndash2720 doi101111j0014-38202003tb01514x

Dettman JR Jacobson DJ Turner E Pringle A TaylorJW (2003b) Reproductive isolation and phylogeneticdivergence in Neurospora comparing methods ofspecies recognition in a model eukaryote Evolution57(12)2721 doi10155403-074

Devkota S Cornejo C Werth S Chaudhary RP Sche-idegger C (2014) Characterization of microsatelliteloci in the Himalayan lichen fungus Lobaria pindar-ensis (Lobariaceae) Appl Plant Sci 2(5)1300101doi103732apps1300101

Divakar PK Molina MC Lumbsch HT Crespo A (2005)Parmelia barrenoae a new lichen species related toParmelia sulcata (Parmeliaceae) based on molecularand morphological data The Lichenologist 37(01)37ndash46 doi101017S0024282904014641

Divakar PK Amo De paz G del Prado R Esslinger TLCrespo A (2007) Upper cortex anatomy corroboratesphylogenetic hypothesis in species of Physconia(Ascomycota Lecanoromycetes) Mycol Res 111(11)1311ndash1320 doi101016jmycres200708009

Divakar PK Figueras G Hladun N Crespo A (2010)Molecular phylogenetic studies reveal an undescribedspecies within the North American concept of Mela-nelixia glabra (Parmeliaceae) Fungal Divers 42(1)47ndash55 doi101007s13225-010-0027-3

Donoghue MJ Gauthier A (2004) Implementing thephylocode Trends Ecol Evol 19(6)281ndash282 doi101016jtree200404004

Eaton DAR Ree RH (2013) Inferring phylogeny andintrogression using RADseq data an example fromflowering plants (Pedicularis Orobanchaceae) SystBiol 62(5)689ndash706 doi101093sysbiosyt032

Edwards SV (2009) Is a new and general theory ofmolecular systematics emerging Evolution 63(1)1ndash19 doi101111j1558-5646200800549x

Edwards DL Knowles LL (2014) Species detection andindividual assignment in species delimitation canintegrative data increase efficacy Proc R Soc B BiolSci 281(1777) doi101098rspb20132765

Egan RS (1986) Correlations and non-correlations ofchemical variation patterns with lichen morphologyand geography The Bryologist 8999ndash110

Elix JA Corush J Lumbsch HT (2009) Triterpenechemosyndromes and subtle morphological characterscharacterise lineages in the Physcia aipolia group inAustralia (Ascomycota) Syst Biodivers 7(04)479ndash487 doi101017S1477200009990223

Emerson KJ Merz CR Catchen JM Hohenlohe PACresko WA Bradshaw WE Holzapfel CM (2010)Resolving postglacial phylogeography using high-throughput sequencing Proc Natl Acad Sci 107(37)16196ndash16200 doi101073pnas1006538107

Ence DD Carstens BC (2011) SpedeSTEM a rapid andaccurate method for species delimitation Mol EcolResour 11(3)473ndash480 doi101111j1755-0998201002947x

EvannoGRegnaut SGoudet J (2005)Detecting thenumberof clusters of individuals using the software STRUC-TURE a simulation study Mol Ecol 14(8)2611ndash2620doi101111j1365-294X200502553x

Falush D Stephens M Pritchard JK (2003) Inference ofpopulation structure using multilocus genotype datalinked loci and correlated allele frequencies Genetics164(4)1567ndash1587 doi101111j1471-8286200701758x

Fan HH Kubatko LS (2011) Estimating species treesusing approximate Bayesian computation Mol Phylo-genet Evol 59(2)354ndash363 doi101016jympev201102019

Fernaacutendez-Mendoza F Printzen C (2013) Pleistoceneexpansion of the bipolar lichen Cetraria aculeata intothe Southern hemisphere Mol Ecol 22(7)1961ndash1983doi101111mec12210

Fernaacutendez-Mendoza F Domaschke S Garciacutea MA JordanP Martin MP Printzen C (2011) Population structure


of mycobionts and photobionts of the widespreadlichen Cetraria aculeata Mol Ecol 20(6)1208ndash1232doi101111j1365-294X201004993x

Fraley C Raftery A (2007) Model-based methods ofclassification Using the mclust software in chemo-metrics J Stat Softw 18i06 [Available online at httpwwwdoajorgdoajfunc=abstractampid=218544]

Fujisawa T Barraclough TG (2013) Delimiting speciesusing single-locus data and the generalized mixed yulecoalescent (GMYC) approach a revised method andevaluation on simulated datasets Syst Biol 62(5)707ndash724 doi101093sysbiosyt033

Fujita MK Leacheacute AD Burbrink FT McGuire JA MoritzC (2012) Coalescent-based species delimitation in anintegrative taxonomy Trends Ecol Evol 27(9)480ndash488 doi101016jtree201204012

Gatesy J OrsquoGrady P Baker RH (1999) Corroborationamong data sets in simultaneous analysis hiddensupport for phylogenetic relationships among higherlevel Artiodactyl taxa Cladistics 15(3)271ndash313doi101111j1096-00311999tb00268x

Gladieux P Ropars J Badouin H Branca A Aguileta Gde Vienne DM Rodriacuteguez de la Vega RC Branco SGiraud T (2014) Fungal evolutionary genomicsprovides insight into the mechanisms of adaptivedivergence in eukaryotes Mol Ecol 23(4)753ndash773doi101111mec12631

Goffinet B Miadlikowska J Goward T (2003) Phyloge-netic inferences based on nrDNA sequences supportfive morphospecies within the Peltigera didactylacomplex (Lichenized Ascomycota) The Bryologist106(3)349ndash364 doi10163901

Gowan SP (1986) Evolution of secondary natural productsin the genus Porpidia (Ascomycata Porpidiaceae) AmJ Bot 73606

Griffin PC Hoffmann AA (2014) Limited genetic diver-gence among Australian alpine Poa tussock grassescoupled with regional structuring points to ongoinggene flow and taxonomic challenges Ann Bot doi101093aobmcu017

Grube M Hawksworth DL (2007) Trouble with lichenthe re-evaluation and re-interpretation of thallus formand fruit body types in the molecular era Mycol Res111(9)1116ndash1132 doi101016jmycres200704008

Gueidan C Savi S Thues H Roux C Keller C Tibell LPrieto M Heimarsson S Breuss O Orange A FrobergL Wynns AA Navarro-Rosines P Krzewicka BPykaelae J Grube M Lutzoni F (2009) Genericclassification of the Verrucariaceae (Ascomycota)based on molecular and morphological evidencerecent progress and remaining challenges Taxon 58(1)184ndash208

Guillot G Mortier F Estoup A (2005) Geneland acomputer package for landscape genetics Mol EcolNotes 5(3)712ndash715 doi101111j1471-8286200501031x

Guillot G Renaud S Ledevin R Michaux J Claude J(2012) A unifying model for the analysis of pheno-typic genetic and geographic data Syst Biol 61(6)897ndash911 doi101093sysbiosys038

Guisan A Thuiller W (2005) Predicting species distribu-tion offering more than simple habitat models EcolLett 8(9)993ndash1009 doi101111j1461-0248200500792x

Hale ME (1990) A synopsis of the lichen genusXanthoparmelia (Vainio) Hale (AscomycotinaParmeliaceae) Smithsonian Institution Press Wash-ington DC

Hamilton CA Hendrixson BE Brewer MS Bond JE(2014) An evaluation of sampling effects on multipleDNA barcoding methods leads to an integrativeapproach for delimiting species a case study ofthe North American tarantula genus Aphonopelma(Araneae Mygalomorphae Theraphosidae) MolPhylogenet Evol 7179ndash93 doi101016jympev201311007

Hausdorf B (2011) Progress toward a general speciesconcept Evolution 65(4)923ndash931 doi101111j1558-5646201101231x

Hausdorf B Hennig C (2010) Species delimitation usingdominant and codominant multilocus markers SystBiol 59(5)491ndash503 doi101093sysbiosyq039

Hawksworth DL (1976) Lichen chemotaxonomy InBailey RH (ed) Lichenology progress and problemsAcademic London pp 139ndash184

Hebert PDN Cywinska A Ball SL deWaard JR (2003)Biological identifications through DNA barcodesProc R Soc Lond B Biol Sci 270(1512)313ndash321doi101098rspb20022218

Hebert PDN Stoeckle MY Zemlak TS Francis CM(2004) Identification of birds through DNA barcodesPLoS Biol 2(10)e312 doi101371journalpbio0020312

Heled J Drummond AJ (2010) Bayesian inference ofspecies trees from multilocus data Mol Biol Evol 27(3)570ndash580 doi101093molbevmsp274

Hey J (2006) On the failure of modern species conceptsTrends Ecol Evol 21(8)447ndash450 doi101016jtree200605011

Hibbett DS OhmanA Glotzer D NuhnMKirk P NilssonRH (2011) Progress in molecular and morphologicaltaxon discovery in Fungi and options for formalclassification on environmental sequences Fungal BiolRev 2538ndash47 doi101016jfbr201101001

Houmlgnabba F Wedin M (2003) Molecular phylogeny ofthe Sphaerophorus globosus species complex Cladis-tics 19(3)224ndash232 doi101111j1096-00312003tb00365x

Hudson RR Coyne JA (2002) Mathematical conse-quences of the genealogical species concept Evolu-tion 56(8)1557 doi101111j0014-38202002tb01467x

Huelsenbeck JP Andolfatto P Huelsenbeck ET (2011)Structurama Bayesian inference of population struc-ture Evol Bioinf Online 7(2011)55ndash59 doi104137EBOS6761

Jones G Oxelman B (2014) DISSECT an assignment-free Bayesian discovery method for species delimita-tion under the multispecies coalescent Bioinformaticsdoi101101003178

38 SD Leavitt et al

Kekkonen M Hebert PD (2014) DNA barcode-baseddelineation of putative species efficient start fortaxonomic workflows Mol Ecol Res doi1011111755-099812233

Kelly LJ Hollingsworth PM Coppins BJ Ellis CJHarrold P Tosh J Yahr R (2011) DNA barcoding oflichenized fungi demonstrates high identification suc-cess in a floristic context New Phytol 191(1)288ndash300doi101111j1469-8137201103677x

Kiss L (2012) Limits of nuclear ribosomal DNA internaltranscribed spacer (ITS) sequences as species barcodesfor fungi Proc Natl Acad Sci 109(27)E1811ndashE1811doi101073pnas1207143109

Kluge AG (1989) A concern for evidence and aphylogenetic hypothesis for relationships among Epi-crates (Boidae Serpentes) Syst Zool 38(1)7ndash25doi101093sysbio3817

Knowles LL Carstens BC (2007) Delimiting specieswithout monophyletic gene trees Syst Biol 56(6)887ndash895 doi10108010635150701701091

Kroken S Taylor JW (2001) A gene genealogical approachto recognize phylogenetic species boundaries in thelichenized fungus Letharia Mycologia 93(1)38ndash53

Kubatko LS Degnan JH (2007) Inconsistency of phylo-genetic estimates from concatenated data under coa-lescence Syst Biol 56(1)17ndash24 doi10108010635150601146041

Latch EK Dharmarajan G Glaubitz JC Rhodes OE Jr(2006) Relative performance of Bayesian clusteringsoftware for inferring population substructure andindividual assignments at low levels of populationdifferentiation Conserv Genet 7(2)295ndash302 doi101007s10592-005-9098-1

Lawrey JD (1986) Biological role of lichen substancesThe Bryologist 89(2)111ndash122

Le Gac M Hood ME Fournier E Giraud T (2007)Phylogenetic evidence of host-specific cryptic speciesin the anther smut fungus Evolution 61(1)15ndash26doi101111j1558-5646200700002x

Leacheacute AD (2009) Species tree discordance traces tophylogeographic clade boundaries in North Americanfence lizards (Sceloporus) Syst Biol 58(6)547ndash559doi101093sysbiosyp057

Leacheacute AD Fujita MK (2010) Bayesian species delim-itation in West African forest geckos (Hemidactylusfasciatus) Proc R Soc B Biol Sci 2773071ndash3077doi101098rspb20100662

Leacheacute AD Koo M Spencer C Papenfuss T Fisher RMcGuire J (2009) Quantifying ecological morpho-logical and genetic variation to delimit species in thecoast horned lizard species complex (Phrynosoma)Proc Natl Acad Sci USA 10612418ndash12423 doi101073pnas0906380106

Leacheacute AD Fujita MK Minin VN Bouckaert RR (2014)Species delimitation using genome-wide SNP dataSystematic Biology doi101093sysbiosyu018

Leavitt SD Fankhauser JD Leavitt DH Porter LDJohnson LA St Clair LL (2011a) Complex patterns ofspeciation in cosmopolitan ldquorock posyrdquo lichensmdashdiscovering and delimiting cryptic fungal species in

the lichen-forming Rhizoplaca melanophthalma spe-cies-complex (Lecanoraceae Ascomycota) MolPhylogenet Evolu 59(3)587ndash602 doi101016jympev201103020

Leavitt SD Johnson L St Clair LL (2011b) Speciesdelimitation and evolution in morphologically andchemically diverse communities of the lichen-forminggenus Xanthoparmelia (Parmeliaceae Ascomycota) inwestern North America Am J Bot 98 (2)175ndash188doi103732ajb1000230

Leavitt SD Johnson LA Goward T St Clair LL (2011c)Species delimitation in taxonomically difficult lichen-forming fungi an example from morphologically andchemically diverse Xanthoparmelia (Parmeliaceae) inNorth America Mol Phylogen Evol 60(3)317ndash332doi101016jympev201105012

Leavitt S Esslinger T Divakar P Lumbsch H (2012a)Miocene and Pliocene dominated diversification of thelichen-forming fungal genus Melanohalea (Parmelia-ceae Ascomycota) and Pleistocene population expan-sions BMC Evol Biol 12(1)176 doi1011861471-2148-12-176

Leavitt SD Esslinger TL Divakar PK Lumbsch HT(2012b) Miocene divergence phenotypically crypticlineages and contrasting distribution patterns incommon lichen-forming fungi (Ascomycota Parmeli-aceae) Biol J Linn Soc 1007920ndash937 doi101111j1095-8312201201978x

Leavitt SD Esslinger TL Lumbsch HT (2012c) Neogene-dominated diversification in neotropical montanelichens dating divergence events in the lichen-form-ing fungal genus Oropogon (Parmeliaceae) Am J Bot99(11)1764ndash1777 doi103732ajb1200146

Leavitt SD Esslinger TL Nelsen MP Lumbsch HT(2013a) Further species diversity in NeotropicalOropogon (Lecanoromycetes Parmeliaceae) in Cen-tral America The Lichenologist 45(04)553ndash564doi101017S0024282913000212

Leavitt SD Esslinger TL Spribille T Divakar PKLumbsch HT (2013b) Multilocus phylogeny of thelichen-forming fungal genus Melanohalea (Parmelia-ceae Ascomycota) insights on diversity distributionsand a comparison of species tree and concatenatedtopologies Mol Phylogenet Evol 66(2013)138ndash152doi101016jympev201209013

Leavitt SD Fernaacutendez-Mendoza F Peacuterez-Ortega SSohrabi M Divakar PK Lumbsch HT St Clair LL(2013c) DNA barcode identification of lichen-formingfungal species in the Rhizoplaca melanophthalmaspecies-complex (Lecanorales Lecanoraceae) includ-ing five new species MycoKeys 71ndash22 doi103897mycokeys74508

Leavitt SD Fernaacutendez-Mendoza F Peacuterez-Ortega SSohrabi M Divakar PK Vondraacutek J Thorsten Lum-bsch H Clair LLS (2013d) Local representation ofglobal diversity in a cosmopolitan lichen-formingfungal species complex (Rhizoplaca Ascomycota)J Biogeogr 40(9)1792ndash1806 doi101111jbi12118

Leavitt SD Lumbsch HT Stenroos S St Clair LL(2013e) Pleistocene speciation in North American


lichenized fungi and the impact of alternative speciescircumscriptions and rates of molecular evolution ondivergence estimates PLoS ONE 8(12)e85240doi101371journalpone0085240

Leavitt SD Esslinger TL Hansen ES Divakar PK CrespoA Loomis BF Lumbsch HT (2014) DNA barcoding ofbrown Parmeliae (Parmeliaceae) species a molecularapproach for accurate specimen identification empha-sizing species in Greenland Organ Divers Evol 14(1)11ndash20 doi101007s13127-013-0147-1

Leliaert F Verbruggen H Vanormelingen P Steen FLopez-Bautista JM Zuccarello GC De Clerck O(2014) DNA-based species delimitation in algae Eur JPhycol 49(2) doi101080096702622014904524

Leuckert C (1985) Probleme der Flechten-Chemotaxon-omiemdashStoffkombinationen und ihre taxonomischeWertung Ber Deut Bot Ges 98401ndash408

Lewis ZA Shiver AL Stiffler N Miller MR Johnson EASelker EU (2007) High-density detection of restric-tion-site-associated DNA markers for rapid mappingof mutated loci in Neurospora Genetics 177(2)1163ndash1171 doi101534genetics107078147

Lindblom L Soslashchting U (2008) Taxonomic revision ofXanthomendoza borealis and Xanthoria mawsonii(Lecanoromycetes Ascomycota) The Lichenologist40(05)399ndash409 doi101017S0024282908007937

Lohtander K Kaumlllersjouml M Roland M Tehler A (2000)The family physciaceae in fennoscandia phylogenyinferred from its sequences Mycologia 92(4)728ndash735

Longton RE (1997) The role of bryophytes and lichens inpolar ecosystems In Woodin SJ Marquiss M (eds)Ecology of Arctic Environments Blackwell ScienceOxford pp 69ndash96 (Special publication No 13)

Luumlcking R (2012) Predicting species richness in tropicallichenized fungi with lsquomodularrsquo combinations ofcharacter states Biodiver Conserv 21(9)2341ndash2360doi101007s10531-011-0217-7

Luumlcking R del Prado R Lumbsch HT Will-Wolf SAptroot A Sipman HJM Umana L Chaves JL (2008)Phytogenetic patterns of morphological and chemicalcharacters and reproductive mode in the Heterodermiaobscurata group in Costa Rica (Ascomycota Physci-aceae) Syst Biodivers 6(1)31ndash41 doi101017S1477200007002629

Lumbsch HT (1994) Die Lecanora subfusca-Gruppe inAustralasien J Hattori Bot Lab 771ndash175

Lumbsch HT (1998a) Taxonomic use of metabolic data inlichen-forming fungi In Frisvad JC Bridge PDArora DK (eds) Chemical Fungal Taxonomy MarcelDekker New York pp 345ndash387

Lumbsch HT (1998b) The use of metabolic data inlichenology at the species and subspecific levels TheLichenologist 30(4ndash5)357ndash367 doi101017S0024282992000380

Lumbsch HT (2002) Analysis of phenolic products inlichens for identification and taxonomy Protocols inlichenology Culturing biochemistry ecophysiologyand use in biomonitoring Springer Berlin

Lumbsch HT Leavitt SD (2011) Goodbye morphologyA paradigm shift in the delimitation of species in

lichenized fungi Fungal Divers 50(1)59ndash72 doi101007s13225-011-0123-z

Lumbsch HT Ahti T Altermann S Amo de Paz GAptroot A Arup U Barcenas Pentildea A Bawingan PABenatti MN Betancourt L Bjoumlrk CR Boonpragob KBrand M Bungartz F Caceres MES Candan MChaves JL Clerc P Common R Coppins BJ CrespoA Dal Forno M Divakar PK Duya MV Elix JAElvebakk A Fankhauser J Farkas E Ferraro LIFischer E Galloway DJ Gaya E Giralt M Goward TGrube M Hafellner J Hernandez JE Herrera-CamposMA Kalb K Kaumlrnefelt I Kantvilas G Killmann DKirika P Knudesn K Komposch H Kondratyuk SLawrey JD Mangold A Marcelli MP McCune BPMichlig A Miranda Gonzalez R Moncada B Naika-tini A Nelsen MP Oslashvstedal DO Palice Z Papong KParnmen S Peacuterez-Ortega S Printzen C Rico VJRivas Plata E Robayo J Rosabal D Ruprecht USalazar Allen N Sancho L Santos de Jesus L SantosVieira T Schultz M Seaward MRD Seacuterusiaux ESchmitt I Sipman HJM Sohrabi M Soslashchting USoslashgaard MZ Sparrius LB Spielmann A Spribille TSutjaritturakan J Thammathaworn A Thell A ThorG Thuumls H Timdal E Truong C Tuumlrk R UmantildeaTenorio L Upreti D van den Boom P Vivas RebueltaM Wedin M Will-Wolf S Wirth V Wirtz N Yahr RYeshitela K Ziemmeck F Wheeler T Luumlcking R(2011) One hundred new species of lichenized fungi asignature of undiscovered global diversity Phytotaxa181ndash127

Mark K Saag L Saag A Thell A Randlane T (2012)Testing morphology-based delimitation of Vulpicidajuniperinus and V tubulosus (Parmeliaceae) usingthree molecular markers The Lichenologist 44(06)757ndash772 doi101017S0024282912000448

Martiacuten MP LaGreca S Lumbsch HT (2003) Molecularphylogeny of Diploschistes inferred from ITSsequence data The Lichenologist 35(01)27ndash32doi101006lich20020427

Masters BC Fan V Ross HA (2011) Species delimitationmdasha geneious plugin for the exploration of speciesboundaries Mol Ecol Res 11(1)154ndash157 doi101111j1755-0998201002896x

Mayden RL (1997) A hierarchy of species concepts thedenouement in the saga of the species problem InClaridge MF Dawah HA Wilson MR (eds) Speciesthe units of biodiversity Chapman amp Hall Londonpp 381ndash424

Mayr E (1963) Animal species and evolution HarvardUniversity Press Cambridge

Mayr E (1970) Populations species and evolutionBelknap Press of Harvard University Press Cambridge

McCune B (2000) Lichen communities as indicators offorest health The Bryologist 103(2)353ndash356 doi1016390007-2745(2000)103[0353LCAIOF]20CO2

McCune B Printzen C (2011) Distribution and climaticniches of the Lecanora varia group in western USABibliotheca Lichenologica 106225ndash234

McDonald T Miadlikowska J Lutzoni F (2003) Thelichen genus Sticta in the Great Smoky Mountains a

40 SD Leavitt et al

phylogenetic study of morphological chemical andmolecular data The Bryologist 106(1)61ndash79 doi1016390007-2745(2003)106[0061TLGSIT]20CO2

McKay BD Mays HL Wu Y Li H Yao C-T Nishiumi IZou F (2013) An empirical comparison of character-based and coalescent-based approaches to speciesdelimitation in a young avian complex Mol Ecol 22(19)4943ndash4957 doi101111mec12446

Miller MR Dunham JP Amores A Cresko WA JohnsonEA (2007) Rapid and cost-effective polymorphismidentification and genotyping using restriction siteassociated DNA (RAD) markers Genome Res 17(2)240ndash248 doi101101gr5681207

Miralles A Vences M (2013) New metrics for compar-ison of taxonomies reveal striking discrepanciesamong species delimitation methods in Madascincuslizards PLoS ONE 8(7)e68242 doi101371journalpone0068242

Mishler BD Brandon RN (1987) Individuality pluralismand the phylogenetic species concept Biol Philos 2397

Molina M Crespo A Blanco O Lumbsch HT Hawks-worth DL (2004) Phylogenetic relationships andspecies concepts in Parmelia s str (Parmeliaceae)inferred from nuclear ITS rDNA and β-tubulinsequences The Lichenologist 36(01)37ndash54 doi101017S0024282904013933

Molina M Del-Prado R Divakar P Saacutenchez-Mata DCrespo A (2011) Another example of cryptic diversityin lichen-forming fungi the new species Parmeliamayi (Ascomycota Parmeliaceae) Organ Divers Evol11(5)331ndash342 doi101007s13127-011-0060-4

Monaghan MT Wild R Elliot M Fujisawa T Balke MInward DJG Lees DC Ranaivosolo R Eggleton PBarraclough TG Vogler AP (2009) Acceleratedspecies inventory on Madagascar using coalescent-based models of species delineation Syst Biol 58(3)298ndash311 doi101093sysbiosyp027

Moncada B Reidy B Luumlcking R (2014) A phylogeneticrevision of Hawaiian Pseudocyphellaria (lichenizedAscomycota Lobariaceae) reveals eight new speciesand a high degree of inferred endemism The Bryol-ogist 117(2)119ndash160 doi httpdxdoiorg1016390007-2745-1172119

Moreau CS (2009) Inferring ant evolution in the age ofmolecular data Myrmecological New 12201

Muggia L Grube M Tretiach M (2008) A combinedmolecular and morphological approach to speciesdelimitation in black-fruited endolithic Caloplacahigh genetic and low morphological diversity MycolRes 112(1)36ndash49 doi101016jmycres200702001

Muggia L Peacuterez-Ortega S Fryday A Spribille T GrubeM (2014) Global assessment of genetic variation andphenotypic plasticity in the lichen-forming speciesTephromela atra Fungal Divers 64(1)233ndash251doi101007s13225-013-0271-4

Nash TH Zavada M (1977) Population studies amongSonoran Desert species of Parmelia subg Xanthop-armelia (Parmeliaceae) Am J Bot 64(6)664ndash669

Nylander W (1866a) Circa novum in studio Lichenumcriterium chemicum Flora 49198ndash201

Nylander W (1866b) Hypochlorite of lime and hydrate ofpotash Two new criteria for the study of lichens(Tanslated and communicated by the Rev W ALeighton) Bot J Linn Soc 9358ndash365

OrsquoBrien HE Miadlikowska J Lutzoni F (2009) Assessingreproductive isolation in highly diverse communitiesof the lichen-forming fugnal genus Peltigera Evolu-tion 63(8)2076ndash2086 doi101111j1558-5646200900685x

OrsquoMeara BC (2010) New heuristic methods for jointspecies delimitation and species tree inference SystBiol 59(1)59ndash73 doi101093sysbiosyp077

OrsquoNeill EM Schwartz R Bullock CT Williams JSShaffer HB Aguilar-Miguel X Parra-Olea G Weis-rock DW (2013) Parallel tagged amplicon sequencingreveals major lineages and phylogenetic structure inthe North American tiger salamander (Ambystomatigrinum) species complex Mol Ecol 22(1)111ndash129doi101111mec12049

Orock EA Leavitt SD Fonge BA St Clair LL LumbschHT (2012) DNA-based identification of lichen-form-ing fungi Can publicly available sequence databasesaid in lichen diversity inventories of Mount Cameroon(West Africa) The Lichenologist 44(6)833ndash839doi101017S0024282912000424

Otaacutelora MAG Martiacutenez I Aragoacuten G Molina MC (2010)Phylogeography and divergence date estimates of alichen species complex with a disjunct distributionpattern Am J Bot 97(2)216ndash223 doi103732ajb0900064

Padial J Castroviejo-Fisher S Kohler J Vila C ChaparroJ De la Riva I (2009) Deciphering the products ofevolution at the species level the need for anintegrative taxonomy Zoolog Scr 38(4)431ndash447doi101111j1463-6409200800381x

Padial J Miralles A De la Riva I Vences M (2010) Theintegrative future of taxonomy Front Zoo 7(1)16doi1011861742-9994-7-16

Papong K Luumlcking R Thammathaworn A BoonpragobK (2009) Four new taxa of Chroodiscus (thelotremoidGraphidaceae) from Southeast Asia The Bryologist112(1)152ndash163 doi1016390007-2745-1121152

Park S-Y Choi J Kim JA Jeong M-H Kim S Lee Y-HHur J-S (2013a) Draft genome sequence of Cladoniamacilenta KoLRI003786 a lichen-forming fungusproducing biruloquinone Genome Announcements 1(5)e00695-13 doi101128genomeA00695-13

Park S-Y Choi J Kim JA Yu N-H Kim S KondratyukSY Lee Y-H Hur J-S (2013b) Draft genome sequenceof lichen-forming fungus Caloplaca flavorubescensStrain KoLRI002931 Genome Announcements 1(4)e00678-13 doi101128genomeA00678-13

Park S-Y Choi J Lee G-W Kim JA Oh S-O Jeong M-HYu N-H Kim S Lee Y-H Hur J-S (2014) Draftgenome sequence of lichen-forming fungus Cladoniametacorallifera Strain KoLRI002260 GenomeAnnouncements 2(1)e01065-13 doi101128genomeA01065-13

Parnmen S Rangsiruji A Mongkolsuk P Boonpragob KNutakki A Lumbsch HT (2012) Using phylogenetic


and coalescent methods to understand the speciesdiversity in the Cladia aggregata complex (Ascomy-cota Lecanorales) PLoS ONE 7(12)e52245 doi101371journalpone0052245

Peacuterez-Ortega S Fernaacutendez-Mendoza F Raggio J VivasM Ascaso C Sancho LG Printzen C de los Riacuteos A(2012) Extreme phenotypic variation in Cetrariaaculeata (lichenized Ascomycota) adaptation or inci-dental modification Ann Bot 109(6)1133ndash1148doi101093aobmcs042

Pino-Bodas R Burgaz A Martiacuten M Lumbsch HT (2011)Phenotypical plasticity and homoplasy complicatespecies delimitation in the Cladonia gracilis group(Cladoniaceae Ascomycota) Organ Divers Evol 11(5)343ndash355 doi101007s13127-011-0062-2

Pino-Bodas R Burgaz AR Martin MP Lumbsch HT(2012a) Species delimitations in the Cladonia cariosagroup (Cladoniaceae Ascomycota) The Lichenologist44(01)121ndash135 doi101017S002428291100065X

Pino-Bodas R Martiacuten M Burgaz A (2012b) Cladoniasubturgida and C iberica (Cladoniaceae) form asingle morphologically and chemically polymorphicspecies Mycol Prog 11(1)269ndash278 doi101007s11557-011-0746-1

Pino-Bodas R Ahti T Stenroos S Martiacuten MP BurgazAR (2013a) Multilocus approach to species recogni-tion in the Cladonia humilis complex (CladoniaceaeAscomycota) Am J Bot 100(4)664ndash678 doi103732ajb1200162

Pino-Bodas R Martiacuten AP Burgaz AR Lumbsch HT(2013b) Species delimitation in Cladonia (Ascomy-cota) a challenge to the DNA barcoding philosophyMol Ecol Resour 13(6)1058ndash1068 doi1011111755-099812086

Pons J Barraclough TG Gomez-Zurita J Cardoso ADuran DP Hazell S Kamoun S Sumlin WD VoglerAP (2006) Sequence-based species delimitation for theDNA taxonomy of undescribed insects Syst Biol 55(4)595ndash609 doi10108010635150600852011

Porada P Weber B Elbert W Poumlschl U Kleidon A(2014) Estimating impacts of lichens and bryophyteson global biogeochemical cycles Global BiogeochemCycles 28(2)71ndash85 doi1010022013GB004705

Pringle A Baker DM Platt JL Wares JP Latgeacute JPTaylor JW (2005) Cryptic speciation in the cosmo-politan and clonal human pathogenic fungus Asper-gillus fumigatus Evolution 59(9)1886ndash1899 doi10155404-2411

Printzen C (2009) Lichen systematics the role ofmorphological and molecular data to reconstructphylogenetic relationships Progress in botany vol71 Springer Berlin pp 233ndash275

Pritchard JK Stephens M Donnelly P (2000) Inference ofpopulation structure using multilocus genotype dataGenetics 155(2)945ndash959

Puillandre N Lambert A Brouillet S Achaz G (2012)ABGD automatic barcode gap discovery for primaryspecies delimitation Mol Ecol 21(8)1864ndash1877doi101111j1365-294X201105239x

Rannala B Yang Z (2003) Bayes estimation of speciesdivergence times and ancestral population sizes usingDNA sequences from multiple loci Genetics1641645ndash1656

Raxworthy CJ Ingram CM Rabibisoa N Pearson RG(2007) Applications of ecological niche modelingfor species delimitation a review and empiricalevaluation using day geckos (Phelsuma) fromMadagascar Syst Biol 56(6)907ndash923 doi10108010635150701775111

Ray J (1686) Historia planarum vol 1 Clark LondonReese Naeligsborg R Ekman S Tibell L (2007) Molecular

phylogeny of the genus Lecania (Ramalinaceaelichenized Ascomycota) Mycol Res 111(5)581ndash591doi101016jmycres200703001

Reid N Carstens B (2012) Phylogenetic estimation errorcan decrease the accuracy of species delimitation aBayesian implementation of the general mixed Yule-coalescent model BMC Evol Biol 12(1)196 doi1011861471-2148-12-196

Rissler L Apodaca J (2007) Adding more ecology intospecies delimitation ecological niche models andphylogeography help define cryptic species in theBlack Salamander (Aneides flavipunctatus) Syst Biol56924ndash942 doi10108010635150701703063

Rivas Plata E Luumlcking R (2013) High diversity ofGraphidaceae (lichenized Ascomycota Ostropales) inAmazonian Peruacute Fungal Divers 58(1)13ndash32 doi101007s13225-012-0172-y

Rivas Plata E Lumbsch HT (2011) Parallel evolution andphenotypic divergence in lichenized fungi a casestudy in the lichen-forming fungal family Graphida-ceae (Ascomycota Lecanoromycetes Ostropales)Mol Phylogenet Evol 61(1)45ndash63 doi101016jympev201104025

Rogers RW (1989) Chemical variation and the speciesconcept in lichenized ascomycetes Bot J Linn Soc101229ndash239

Ross HA (2014) The incidence of species-level paraphylyin animals a re-assessment Mol Phylogenet Evol7610ndash17 doi101016jympev201402021

Ross KG Gotzek D Ascunce MS Shoemaker DD (2010)Species delimitation a case study in a problematic anttaxon Syst Biol 59(2)162ndash184 doi101093sysbiosyp089

Rubin BER Ree RH Moreau CS (2012) Inferringphylogenies from RAD sequence data PLoS ONE 7(4)e33394 doi101371journalpone0033394

Rubinoff D (2006) Utility of mitochondrial DNA barcodesin species conservation Conserv Biol 20(4)1026ndash1033 doi101111j1523-1739200600372x

Ruiz-Sanchez E Sosa V (2010) Delimiting speciesboundaries within the Neotropical bamboo Otatea(Poaceae Bambusoideae) using molecular morpho-logical and ecological data Mol Phylogenet Evol 54(2)344ndash356 doi101016jympev200910035

Ruprecht U Lumbsch HT Brunauer G Green TGA TuumlrkR (2010) Diversity of Lecidea (Lecideaceae Asco-mycota) species revealed by molecular data and

42 SD Leavitt et al

morphological characters Antarctic Sci 22 (SpecialIssue 06)727ndash741 doi101017S0954102010000477

Salicini I Ibaacutentildeez C Juste J (2011) Multilocus phylogenyand species delimitation within the Nattererrsquos batspecies complex in the Western Palearctic MolPhylogenet Evol 61(3)888ndash898 doi101016jympev201108010

Satler JD Carstens BC Hedin M (2013) Multilocusspecies delimitation in a complex of morphologicallyconserved trapdoor spiders (Mygalomorphae Antro-diaetidae Aliatypus) Syst Biol 62(6)805ndash823 doi101093sysbiosyt041

Schlick-Steiner BC Steiner FM Seifert B Stauffer CChristian E Crozier RH (2010) Integrative taxonomya multisource approach to exploring biodiversityAnnu Rev Entomol 55(1)421ndash438 doi101146annurev-ento-112408-085432

Schoch CL Seifert KA Huhndorf S Robert V SpougeJL Levesque CA Chen W Fungal Barcoding Con-sortium (2012) Nuclear ribosomal internal transcribedspacer (ITS) region as a universal DNA barcodemarker for Fungi Proc Natl Acad Sci 109(16)6241ndash6246 doi101073pnas1117018109

Seacuterusiaux E Villarreal AJC Wheeler T Goffinet B(2011) Recent origin active speciation and dispersalfor the lichen genus Nephroma (Peltigerales) inMacaronesia J Biogeogr 38(6)1138ndash1151 doi101111j1365-2699201002469x

Shrestha G Peterson SL St Clair LL (2012) Predictingthe distribution of the air pollution sensitive lichenspecies Usnea hirta The Lichenologist 44(04)511ndash521 doi101017S0024282912000060

Simpson GG (1951) The species concept Evolution5285ndash298

Sites JW Marshall JC (2003) Delimiting species a renais-sance issue in systematic biology Trends Ecol Evol18462ndash470 doi101016S0169-5347(03)00184-8

Sites JW Marshall JC (2004) Operational criteria fordelimiting species Annu Rev Ecol Evol Syst 35(1)199ndash227 doi101146annurevecolsys35112202130128

Šlapeta J Loacutepez-Garciacutea P Moreira D (2006) Globaldispersal and ancient cryptic species in the smallestmarine eukaryotes Mol Biol Evol 23(1)23ndash29doi101093molbevmsj001

Spribille T Klug B Mayrhofer H (2011) A phylogeneticanalysis of the boreal lichen Mycoblastus sanguina-rius (Mycoblastaceae lichenized Ascomycota) revealscryptic clades correlated with fatty acid profiles MolPhylogenet Evol 59(3)603ndash614 doi101016jympev201103021

Stenroos SK DePriest PT (1998) SSU rDNA phylogenyof cladoniiform lichens Am J Bot 85(11)1548ndash1559

Talavera G Dincă V Vila R (2013) Factors affectingspecies delimitations with the GMYC model insightsfrom a butterfly survey Methods Ecol Evol 4(12)1101ndash1110 doi1011112041-210X12107

Taylor JW Jacobson DJ Kroken S Kasuga T GeiserDM Hibbett DS Fisher MC (2000) Phylogeneticspecies recognition and species concepts in fungi

Fungal Genet Biol 31(1)21ndash32 doi101006fgbi20001228

Tewksbury J Anderson JGT Bakker JD Billo TJDunwiddie PW Groom MJ Hampton SE HermanSG Levey DJ Machnicki NJ del Rio CM PowerME Rowell K Salomon AK Stacey L TrombulakSC Wheeler TA (2014) Natural historyrsquos place inscience and society BioScience doi101093bioscibiu032

Thell A Houmlgnabba F Elix JA Feuerer T Kaumlrnefelt IMyllys L Randlane T Saag A Stenroos S Ahti TSeaward MRD (2009) Phylogeny of the cetrarioidcore (Parmeliaceae) based on five genetic markersThe Lichenologist 41(05)489ndash511 doi101017S0024282909990090

Velmala S Myllys L Halonen P Goward T Ahti T(2009) Molecular data show that Bryoria fremontiiand B tortuosa (Parmeliaceae) are conspecific TheLichenologist 41(03)231ndash242 doi101017S0024282909008573

Vondraacutek J Řiacuteha P Arup U Soslashchting U (2009) Thetaxonomy of the Caloplaca citrina group (Teloschist-aceae) in the Black Sea region with contributions tothe cryptic species concept in lichenology TheLichenologist 41(06)571ndash604 doi101017S0024282909008317

Wagner CE Keller I Wittwer S Selz OM Mwaiko SGreuter L Sivasundar A Seehausen O (2013)Genome-wide RAD sequence data provide unprece-dented resolution of species boundaries and relation-ships in the Lake Victoria cichlid adaptive radiationMol Ecol 22(3)787ndash798 doi101111mec12023

Wang Y-Y Liu B Zhang X-Y Zhou Q-M Zhang T LiH Yu Y-F Zhang X-L Hao X-Y Wang M Wang LWei J-C (2014) Genome characteristics reveal theimpact of lichenization on lichen-forming fungusEndocarpon pusillum Hedwig (Verrucariales Asco-mycota) BMC Genom 15(1)34 doi1011861471-2164-15-34

Wedin M Westberg M Crewe AT Tehler A Purvis OW(2009) Species delimitation and evolution of metalbioaccumulation in the lichenized Acarospora smar-agdula (Ascomycota Fungi) complex Cladistics 25(2)161ndash172 doi101111j1096-0031200900240x

Weisrock DW Rasoloarison RM Fiorentino I RalisonJM Goodman SM Kappeler PM Yoder AD (2010)Delimiting species without nuclear monophyly inMadagascarrsquos mouse lemurs PLoS ONE 5(3)e9883doi101371journalpone0009883

Werth S Cornejo C Scheidegger C (2013) Characteriza-tion of microsatellite loci in the lichen fungus Lobariapulmonaria (Lobariaceae) Appl Plant Sci 1(2)1200290 doi103732apps1200290

Westberg M Arup U Kaumlrnefelt I (2007) Phylogeneticstudies in the Candelariaceae (lichenized Ascomy-cota) based on nuclear ITS DNA sequence dataMycol Res 111(11)1277ndash1284 doi101017S0953756204002102

Wiemers M Fiedler K (2007) Does the DNA barcodinggap existmdashA case study in blue butterflies


(Lepidoptera Lycaenidae) Front Zoo 4(1)8 doi1011861742-9994-4-8

Wiens JJ (1998) Combining data sets with differentphylogenetic histories Syst Biol 4568ndash581 doi101080106351598260581

Wiens JJ Penkrot T (2002) Delimiting species usingDNA and morphological variation and discordantspecies limits in spiny lizards (Sceloporus) Syst Biol5169ndash91 doi101080106351502753475880

Wiens JJ Servedio MR (2000) Species delimitation insystematics inferring diagnostic differences betweenspecies Proc R Soc Lond B Biol Sci 267(1444)631ndash636 doi101098rspb20001049

Will KW Mishler BD Wheeler QD (2005) The perils ofDNA barcoding and the need for integrative taxon-omy Syst Biol 54(5)844ndash851 doi10108010635150500354878

Wirtz N Printzen C Lumbsch HT (2012) Using haplo-type networks estimation of gene flow and phenotypic

characters to understand species delimitation in fungiof a predominantly Antarctic Usnea group (Ascomy-cota Parmeliaceae) Organ Divers Evol 12(1)17ndash37doi101007s13127-011-0066-y

Yang Z Rannala B (2010) Bayesian species delimitationusing multilocus sequence data Proc Natl Acad Sci107(20)9264ndash9269 doi101073pnas0913022107

Yeates DK Seago A Nelson L Cameron SL JosephLEO Trueman JWH (2011) Integrative taxonomy oriterative taxonomy Syst Entomol 36(2)209ndash217doi101111j1365-3113201000558x

Zhang C Zhang D-X Zhu T Yang Z (2011) Evaluationof a Bayesian coalescent method of species delimita-tion Syst Biol 60(6)747ndash761 doi101093sysbiosyr071

Zhang J Kapli P Pavlidis P Stamatakis A (2013) Ageneral species delimitation method with applicationsto phylogenetic placements Bioinformatics 29(22)2869ndash2876 doi101093bioinformaticsbtt499

44 SD Leavitt et al

2 The Dynamic Discipline of Species Delimitation Progress Toward Effectively Recognizing Species Boundaries in Natural Populations

Abstract

21 Introduction Whats in a Name The Importance of Accurate Species Delimitations


212 Species in Lichenized Fungi Cases of Cryptic Diversity Polymorphic Lineages and Striking Biogeographic Patterns

22 A Practical Guide to Contemporary Species Delimitation

221 Corroborating Traditional Taxonomy and Discovering Cryptic Species Using Single-Locus Data

222 Sampling Across the Genome Multilocus Sequence Data and Genome-Wide Markers

23 Can We Make Species Delimitation in Lichen-Forming Fungi Truly Integrative


24 Conclusions What About Taxonomy

Acknowledgments

References

21 Introduction Whatrsquosin a Name The Importanceof Accurate SpeciesDelimitations

Although there are over 15 million species for-mally named by scientists current estimates ofthe number of species alive on the planet todayrange from approximately two million to overone hundred million (Caley et al 2014) Docu-menting describing and naming this diversity isparamount for conservation human health foodsecurity and recreation (Tewksbury et al 2014)In a broad sense species delimitation is theprocess of identifying how individuals andpopulations fit into natural species-level clustersand not simply constructs of classification(Carstens et al 2013) Empirical species delimi-tation is a dynamic discipline with ongoingmethodological and bioinformatical develop-ments due to a growing interest in empiricalspecies delimitations Novel analytical methodsincreasing availability of geneticgenomic dataincreasing computation power reassessments ofmorphological and chemical characters andimproved availability of distributional and eco-logical records offer exciting avenues for empir-ically exploring species delimitation andevolutionary relationships in species all over theworld In this chapter we aim to contribute acontemporary perspective on delimiting speciesand offer practical direction for empirical speciesdelimitation studies

To illustrate the importance and difficulties indocumenting and describing biodiversity we willuse the lichens as an example Lichens describe amutualism between a fungus (mycobiont) and aphotosynthetic partner (photobiont) which can beeither a green algae andor cyanobacteriumLichens are ubiquitous components of most ter-restrial ecosystems playing important ecologicalroles and contributing to global biogeochemicalcycles (Porada et al 2014 Bonan and Shugart1989 Longton 1997) Due to the fact that manylichens live and grow continuously for decades oreven hundreds of years showing cumulativeresponses to changes in atmospheric pollutionlevels landmanagement practices and fluctuation

climate these are commonly used as bioindicatorsto monitor the impacts of air pollution forest agesoil quality and climate change (McCune 2000)As iconic examples of symbiosis lichens alsoprovide crucial insight into general patterns andprocesses in symbiotic systems Central tounderstanding the dynamic roles of lichens is ourability to accurately delimit and recognize speciesboundaries Increased accuracy in recognizingspecies boundaries in lichenized fungi has majorimplications for enhancing our perspective onbiological diversity evolution ecology symbi-otic interactions biomonitoring research andconservation policy


Species serve as a central unit for categorizingbiological diversity Humans including bird-watchers doctors fisherman gardeners politi-cians scientists and others rely to varyingdegrees on recognizing species for distinguishingdifferent kinds of organisms and effective com-munication In the biological sciences speciesrepresent one of the most fundamental units andprovide a valuable context for organizing eval-uating and communicating important biologicalconcepts and principles (Coyne and Orr 2004Mayr 1963 Darwin 1859) Due to the fact thatbiological information is commonly providedwith reference to a species unit accurate speciescircumscriptions are integral to interpreting bio-logical patterns and processes across a widerange of subdisciplines in biology (eg anatomybehavior ecology evolution physiology etc)

In spite of the underlying importance of spe-cies in biology the conceptualization of the termldquospeciesrdquo remains somewhat contentious (deQueiroz 2007 Hausdorf 2011 Coyne and Orr2004 Mayden 1997 Simpson 1951 Mayr1963) Most biologists agree that biological nat-ure is aggregated into discrete evolutionarilyindependent entities ie ldquospeciesrdquo (Coyne andOrr 2004) although theorists and empiricistsalike continue to debate over an all-encompass-ing species concept and appropriate operationalcriteria for delimiting species (Hausdorf 2011

12 SD Leavitt et al














14 SD Leavitt et al












16 SD Leavitt et al








Table

21

Somemetho

dsused

forspeciesdelim

itatio

ninclud

ingspeciesdiscov

erymetho

dsforassign

ingsamples

topo

pulatio

nswith

outapriori

inform

ation(BAPS

Gaussianclustering

Guillo

trsquosUnified

Mod

elS

TRUCTURES

TRUCTURAMA)genetic

distance-based

metho

dforsortingsequ

encesinto

hypo

theticalspecies(A

BGD)tree-

basedspeciesdiscov

erymetho

ds(bGMYCG

MYCb

PTP

PTP

ldquoSpecies

Delim

itatio

nrdquoplug

-inforGeneiou

s)and

jointdiscov

eryandvalid

ationmetho

ds(BPamp

PBrownie

DISSE

CTSp

eDeStem)

Metho

dDescriptio

nInpu

tdata

BAPS

mdashpo

pulatio

nassignmentusing

Bayesianclustering

Aprogram


thegenetic

structurein

apopulatio

nclustering

molecular

data


analysesGenetic

mixture

analyses

canbe

performed

atbo

thgroupandindividu

allevelsusingeither

ano

n-spatialo

rspatialm

odel

BAPS

treatsboth

theallele

frequenciesof

themolecular

markers

(ornucleotid

efrequenciesforDNA

sequ

ence

data)andthenu

mberof

genetically

diverged

groups

inpopulatio

nas

random

variablesIn


linkedlocirdquomod

elagenetic

mixture

analysiscanbe

done

usinghaploidsequ

ence

data

orotherlin

kedgenetic

markersA

nalysesandmodel

comparisons

canalso

beperformed

usingafixednu

mber

ofgenetically

diverged

grou

psor

prespecified

populatio

nstructures

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Availablefrom

httpwwwhelsinkifi

bsgsoftwareBAPS

describedin

Coran-

deret

al(200

420

0620


06)

Genotyp


dataor

linkedmarkers

(AFL

Por

SNPs)


pulatio

nassignment


Aprogram


thegenetic

structurein

apopulatio

nModelgroups

sampleinto

popu

latio

nsusinggeno

typicdata


that

canbe

attributed

tobeingmixturesof

norm


nsG

aussianclustering

requires

adatasetwhere


variable


used

with


nto

integrated

datasets

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Implem

entedin

Rusingtheprabclus


10)andmclust

packages

(FraleyandRaftery

2007)

Genotyp

icdata

(flexible)

(con

tinued)

18 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Guillo

trsquosUnified

Mod

elmdashpo

pulatio


Thisapproach

provides

astatistical

model

that

allowsoneto

analyzegenetic

and

phenotyp

ticdata

with

inaun

ified

mod

elandinferencefram

eworkandop

tionally

incorporateinform


nof

samplesA

Bayesianclustering

algorithm

assumes

that

each

clusterin

ageographical

domaincanbe

approxim

ated

bypo

lygons

that

arecentered

around

pointsgeneratedby

aPo

issonprocessGuillo

trsquos

Unified

Mod

elisflexiblein

term

sof

thegenetic

data

that

itcanutilize

andcapableof

accurately

delim

iting

species

Limita

tions

Genetic

andphenotypic

data


historiesfor


divergence

forneutralgenetic

markers

andadaptatio

nfora

morphological

structure

Source

Availableas

anextensionof

theRGENELAND

package(G

uillo

tet

al20

05

2012)(http

www2im

mdtudk

gigu

Geneland)

Genotyp

icandno

n-genetic

(eg

phenotypicalecolog

icalgeograph

ical

behavioral)data

STRUCTUREmdash

populatio

nassignment


Amodel-based

clustering


genotype

data

toinferpopulatio

nstructureandassign

individualsto

populatio

nsIndividualsin


probabilistically

topopulatio

nso

rjointly

totwoor

morepopulatio

nsiftheirgenotypes

indicatethatthey

areadmixedT

hemodeldoes

notassum

eaparticular

mutationprocess

anditcanbe

appliedto

mosto

fthecommonly

used

genetic

markersp

rovidedthat

they

areno

tcloselylin

ked

The


lyaccurate

assignmentsusing

modestn

umbersof

loci(Pritchard

etal2

000)T


fpopulatio

nstructurecanbe

inferred

byassessinglik

elihoodscores

orthead

hocΔKstatistic

(Evann

oet

al20

05)

Limita

tions

Identifying


genetic

clusters

ischallenging

clusters

produced

byST

RUCTUREcanbe

strongly

influenced

byvariationin

sample

sizeclusters

createdby

STRUCTUREmay

notbe

consistent

with

theevolutionary

historyof

thepopulatio


elylong

divergence

times

with

inevolutionary

lineagesTem

poraldivergence

andrelatio

nships

amongputativ

egroups


atedEquivalence

togenetic

clusters


isuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source


ldescribedin

Falush

etal(200

3)andPritchard

etal(200

0)

Genotyp

icdata

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim


ergdisequ

ilibrium

foragiven


everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition


ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram


Huelsenbeck

etal(201

1)

Genotyp

icdata


distances

ldquoAutom

aticBarcode

Gap



thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto


should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model


aYulemod


which

isthen

fitto

aninferred

sing

le-gene

topology

toestim


andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san


requires

anultram

etricgene

tree

andrecent


phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum



nthepresence

ofahigh


numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim


ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto


site

noahmreidhom


Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References














14 SD Leavitt et al












16 SD Leavitt et al








Table

21

Somemetho

dsused

forspeciesdelim

itatio

ninclud

ingspeciesdiscov

erymetho

dsforassign

ingsamples

topo

pulatio

nswith

outapriori

inform

ation(BAPS

Gaussianclustering

Guillo

trsquosUnified

Mod

elS

TRUCTURES

TRUCTURAMA)genetic

distance-based

metho

dforsortingsequ

encesinto

hypo

theticalspecies(A

BGD)tree-

basedspeciesdiscov

erymetho

ds(bGMYCG

MYCb

PTP

PTP

ldquoSpecies

Delim

itatio

nrdquoplug

-inforGeneiou

s)and

jointdiscov

eryandvalid

ationmetho

ds(BPamp

PBrownie

DISSE

CTSp

eDeStem)

Metho

dDescriptio

nInpu

tdata

BAPS

mdashpo

pulatio

nassignmentusing

Bayesianclustering

Aprogram


thegenetic

structurein

apopulatio

nclustering

molecular

data


analysesGenetic

mixture

analyses

canbe

performed

atbo

thgroupandindividu

allevelsusingeither

ano

n-spatialo

rspatialm

odel

BAPS

treatsboth

theallele

frequenciesof

themolecular

markers

(ornucleotid

efrequenciesforDNA

sequ

ence

data)andthenu

mberof

genetically

diverged

groups

inpopulatio

nas

random

variablesIn


linkedlocirdquomod

elagenetic

mixture

analysiscanbe

done

usinghaploidsequ

ence

data

orotherlin

kedgenetic

markersA

nalysesandmodel

comparisons

canalso

beperformed

usingafixednu

mber

ofgenetically

diverged

grou

psor

prespecified

populatio

nstructures

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Availablefrom

httpwwwhelsinkifi

bsgsoftwareBAPS

describedin

Coran-

deret

al(200

420

0620


06)

Genotyp


dataor

linkedmarkers

(AFL

Por

SNPs)


pulatio

nassignment


Aprogram


thegenetic

structurein

apopulatio

nModelgroups

sampleinto

popu

latio

nsusinggeno

typicdata


that

canbe

attributed

tobeingmixturesof

norm


nsG

aussianclustering

requires

adatasetwhere


variable


used

with


nto

integrated

datasets

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Implem

entedin

Rusingtheprabclus


10)andmclust

packages

(FraleyandRaftery

2007)

Genotyp

icdata

(flexible)

(con

tinued)

18 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Guillo

trsquosUnified

Mod

elmdashpo

pulatio


Thisapproach

provides

astatistical

model

that

allowsoneto

analyzegenetic

and

phenotyp

ticdata

with

inaun

ified

mod

elandinferencefram

eworkandop

tionally

incorporateinform


nof

samplesA

Bayesianclustering

algorithm

assumes

that

each

clusterin

ageographical

domaincanbe

approxim

ated

bypo

lygons

that

arecentered

around

pointsgeneratedby

aPo

issonprocessGuillo

trsquos

Unified

Mod

elisflexiblein

term

sof

thegenetic

data

that

itcanutilize

andcapableof

accurately

delim

iting

species

Limita

tions

Genetic

andphenotypic

data


historiesfor


divergence

forneutralgenetic

markers

andadaptatio

nfora

morphological

structure

Source

Availableas

anextensionof

theRGENELAND

package(G

uillo

tet

al20

05

2012)(http

www2im

mdtudk

gigu

Geneland)

Genotyp

icandno

n-genetic

(eg

phenotypicalecolog

icalgeograph

ical

behavioral)data

STRUCTUREmdash

populatio

nassignment


Amodel-based

clustering


genotype

data

toinferpopulatio

nstructureandassign

individualsto

populatio

nsIndividualsin


probabilistically

topopulatio

nso

rjointly

totwoor

morepopulatio

nsiftheirgenotypes

indicatethatthey

areadmixedT

hemodeldoes

notassum

eaparticular

mutationprocess

anditcanbe

appliedto

mosto

fthecommonly

used

genetic

markersp

rovidedthat

they

areno

tcloselylin

ked

The


lyaccurate

assignmentsusing

modestn

umbersof

loci(Pritchard

etal2

000)T


fpopulatio

nstructurecanbe

inferred

byassessinglik

elihoodscores

orthead

hocΔKstatistic

(Evann

oet

al20

05)

Limita

tions

Identifying


genetic

clusters

ischallenging

clusters

produced

byST

RUCTUREcanbe

strongly

influenced

byvariationin

sample

sizeclusters

createdby

STRUCTUREmay

notbe

consistent

with

theevolutionary

historyof

thepopulatio


elylong

divergence

times

with

inevolutionary

lineagesTem

poraldivergence

andrelatio

nships

amongputativ

egroups


atedEquivalence

togenetic

clusters


isuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source


ldescribedin

Falush

etal(200

3)andPritchard

etal(200

0)

Genotyp

icdata

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim


ergdisequ

ilibrium

foragiven


everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition


ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram


Huelsenbeck

etal(201

1)

Genotyp

icdata


distances

ldquoAutom

aticBarcode

Gap



thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto


should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model


aYulemod


which

isthen

fitto

aninferred

sing

le-gene

topology

toestim


andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san


requires

anultram

etricgene

tree

andrecent


phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum



nthepresence

ofahigh


numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim


ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto


site

noahmreidhom


Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References












16 SD Leavitt et al








Table

21

Somemetho

dsused

forspeciesdelim

itatio

ninclud

ingspeciesdiscov

erymetho

dsforassign

ingsamples

topo

pulatio

nswith

outapriori

inform

ation(BAPS

Gaussianclustering

Guillo

trsquosUnified

Mod

elS

TRUCTURES

TRUCTURAMA)genetic

distance-based

metho

dforsortingsequ

encesinto

hypo

theticalspecies(A

BGD)tree-

basedspeciesdiscov

erymetho

ds(bGMYCG

MYCb

PTP

PTP

ldquoSpecies

Delim

itatio

nrdquoplug

-inforGeneiou

s)and

jointdiscov

eryandvalid

ationmetho

ds(BPamp

PBrownie

DISSE

CTSp

eDeStem)

Metho

dDescriptio

nInpu

tdata

BAPS

mdashpo

pulatio

nassignmentusing

Bayesianclustering

Aprogram


thegenetic

structurein

apopulatio

nclustering

molecular

data


analysesGenetic

mixture

analyses

canbe

performed

atbo

thgroupandindividu

allevelsusingeither

ano

n-spatialo

rspatialm

odel

BAPS

treatsboth

theallele

frequenciesof

themolecular

markers

(ornucleotid

efrequenciesforDNA

sequ

ence

data)andthenu

mberof

genetically

diverged

groups

inpopulatio

nas

random

variablesIn


linkedlocirdquomod

elagenetic

mixture

analysiscanbe

done

usinghaploidsequ

ence

data

orotherlin

kedgenetic

markersA

nalysesandmodel

comparisons

canalso

beperformed

usingafixednu

mber

ofgenetically

diverged

grou

psor

prespecified

populatio

nstructures

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Availablefrom

httpwwwhelsinkifi

bsgsoftwareBAPS

describedin

Coran-

deret

al(200

420

0620


06)

Genotyp


dataor

linkedmarkers

(AFL

Por

SNPs)


pulatio

nassignment


Aprogram


thegenetic

structurein

apopulatio

nModelgroups

sampleinto

popu

latio

nsusinggeno

typicdata


that

canbe

attributed

tobeingmixturesof

norm


nsG

aussianclustering

requires

adatasetwhere


variable


used

with


nto

integrated

datasets

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Implem

entedin

Rusingtheprabclus


10)andmclust

packages

(FraleyandRaftery

2007)

Genotyp

icdata

(flexible)

(con

tinued)

18 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Guillo

trsquosUnified

Mod

elmdashpo

pulatio


Thisapproach

provides

astatistical

model

that

allowsoneto

analyzegenetic

and

phenotyp

ticdata

with

inaun

ified

mod

elandinferencefram

eworkandop

tionally

incorporateinform


nof

samplesA

Bayesianclustering

algorithm

assumes

that

each

clusterin

ageographical

domaincanbe

approxim

ated

bypo

lygons

that

arecentered

around

pointsgeneratedby

aPo

issonprocessGuillo

trsquos

Unified

Mod

elisflexiblein

term

sof

thegenetic

data

that

itcanutilize

andcapableof

accurately

delim

iting

species

Limita

tions

Genetic

andphenotypic

data


historiesfor


divergence

forneutralgenetic

markers

andadaptatio

nfora

morphological

structure

Source

Availableas

anextensionof

theRGENELAND

package(G

uillo

tet

al20

05

2012)(http

www2im

mdtudk

gigu

Geneland)

Genotyp

icandno

n-genetic

(eg

phenotypicalecolog

icalgeograph

ical

behavioral)data

STRUCTUREmdash

populatio

nassignment


Amodel-based

clustering


genotype

data

toinferpopulatio

nstructureandassign

individualsto

populatio

nsIndividualsin


probabilistically

topopulatio

nso

rjointly

totwoor

morepopulatio

nsiftheirgenotypes

indicatethatthey

areadmixedT

hemodeldoes

notassum

eaparticular

mutationprocess

anditcanbe

appliedto

mosto

fthecommonly

used

genetic

markersp

rovidedthat

they

areno

tcloselylin

ked

The


lyaccurate

assignmentsusing

modestn

umbersof

loci(Pritchard

etal2

000)T


fpopulatio

nstructurecanbe

inferred

byassessinglik

elihoodscores

orthead

hocΔKstatistic

(Evann

oet

al20

05)

Limita

tions

Identifying


genetic

clusters

ischallenging

clusters

produced

byST

RUCTUREcanbe

strongly

influenced

byvariationin

sample

sizeclusters

createdby

STRUCTUREmay

notbe

consistent

with

theevolutionary

historyof

thepopulatio


elylong

divergence

times

with

inevolutionary

lineagesTem

poraldivergence

andrelatio

nships

amongputativ

egroups


atedEquivalence

togenetic

clusters


isuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source


ldescribedin

Falush

etal(200

3)andPritchard

etal(200

0)

Genotyp

icdata

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim


ergdisequ

ilibrium

foragiven


everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition


ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram


Huelsenbeck

etal(201

1)

Genotyp

icdata


distances

ldquoAutom

aticBarcode

Gap



thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto


should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model


aYulemod


which

isthen

fitto

aninferred

sing

le-gene

topology

toestim


andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san


requires

anultram

etricgene

tree

andrecent


phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum



nthepresence

ofahigh


numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim


ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto


site

noahmreidhom


Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References








Table

21

Somemetho

dsused

forspeciesdelim

itatio

ninclud

ingspeciesdiscov

erymetho

dsforassign

ingsamples

topo

pulatio

nswith

outapriori

inform

ation(BAPS

Gaussianclustering

Guillo

trsquosUnified

Mod

elS

TRUCTURES

TRUCTURAMA)genetic

distance-based

metho

dforsortingsequ

encesinto

hypo

theticalspecies(A

BGD)tree-

basedspeciesdiscov

erymetho

ds(bGMYCG

MYCb

PTP

PTP

ldquoSpecies

Delim

itatio

nrdquoplug

-inforGeneiou

s)and

jointdiscov

eryandvalid

ationmetho

ds(BPamp

PBrownie

DISSE

CTSp

eDeStem)

Metho

dDescriptio

nInpu

tdata

BAPS

mdashpo

pulatio

nassignmentusing

Bayesianclustering

Aprogram


thegenetic

structurein

apopulatio

nclustering

molecular

data


analysesGenetic

mixture

analyses

canbe

performed

atbo

thgroupandindividu

allevelsusingeither

ano

n-spatialo

rspatialm

odel

BAPS

treatsboth

theallele

frequenciesof

themolecular

markers

(ornucleotid

efrequenciesforDNA

sequ

ence

data)andthenu

mberof

genetically

diverged

groups

inpopulatio

nas

random

variablesIn


linkedlocirdquomod

elagenetic

mixture

analysiscanbe

done

usinghaploidsequ

ence

data

orotherlin

kedgenetic

markersA

nalysesandmodel

comparisons

canalso

beperformed

usingafixednu

mber

ofgenetically

diverged

grou

psor

prespecified

populatio

nstructures

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Availablefrom

httpwwwhelsinkifi

bsgsoftwareBAPS

describedin

Coran-

deret

al(200

420

0620


06)

Genotyp


dataor

linkedmarkers

(AFL

Por

SNPs)


pulatio

nassignment


Aprogram


thegenetic

structurein

apopulatio

nModelgroups

sampleinto

popu

latio

nsusinggeno

typicdata


that

canbe

attributed

tobeingmixturesof

norm


nsG

aussianclustering

requires

adatasetwhere


variable


used

with


nto

integrated

datasets

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

arenot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

Implem

entedin

Rusingtheprabclus


10)andmclust

packages

(FraleyandRaftery

2007)

Genotyp

icdata

(flexible)

(con

tinued)

18 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Guillo

trsquosUnified

Mod

elmdashpo

pulatio


Thisapproach

provides

astatistical

model

that

allowsoneto

analyzegenetic

and

phenotyp

ticdata

with

inaun

ified

mod

elandinferencefram

eworkandop

tionally

incorporateinform


nof

samplesA

Bayesianclustering

algorithm

assumes

that

each

clusterin

ageographical

domaincanbe

approxim

ated

bypo

lygons

that

arecentered

around

pointsgeneratedby

aPo

issonprocessGuillo

trsquos

Unified

Mod

elisflexiblein

term

sof

thegenetic

data

that

itcanutilize

andcapableof

accurately

delim

iting

species

Limita

tions

Genetic

andphenotypic

data


historiesfor


divergence

forneutralgenetic

markers

andadaptatio

nfora

morphological

structure

Source

Availableas

anextensionof

theRGENELAND

package(G

uillo

tet

al20

05

2012)(http

www2im

mdtudk

gigu

Geneland)

Genotyp

icandno

n-genetic

(eg

phenotypicalecolog

icalgeograph

ical

behavioral)data

STRUCTUREmdash

populatio

nassignment


Amodel-based

clustering


genotype

data

toinferpopulatio

nstructureandassign

individualsto

populatio

nsIndividualsin


probabilistically

topopulatio

nso

rjointly

totwoor

morepopulatio

nsiftheirgenotypes

indicatethatthey

areadmixedT

hemodeldoes

notassum

eaparticular

mutationprocess

anditcanbe

appliedto

mosto

fthecommonly

used

genetic

markersp

rovidedthat

they

areno

tcloselylin

ked

The


lyaccurate

assignmentsusing

modestn

umbersof

loci(Pritchard

etal2

000)T


fpopulatio

nstructurecanbe

inferred

byassessinglik

elihoodscores

orthead

hocΔKstatistic

(Evann

oet

al20

05)

Limita

tions

Identifying


genetic

clusters

ischallenging

clusters

produced

byST

RUCTUREcanbe

strongly

influenced

byvariationin

sample

sizeclusters

createdby

STRUCTUREmay

notbe

consistent

with

theevolutionary

historyof

thepopulatio


elylong

divergence

times

with

inevolutionary

lineagesTem

poraldivergence

andrelatio

nships

amongputativ

egroups


atedEquivalence

togenetic

clusters


isuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source


ldescribedin

Falush

etal(200

3)andPritchard

etal(200

0)

Genotyp

icdata

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim


ergdisequ

ilibrium

foragiven


everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition


ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram


Huelsenbeck

etal(201

1)

Genotyp

icdata


distances

ldquoAutom

aticBarcode

Gap



thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto


should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model


aYulemod


which

isthen

fitto

aninferred

sing

le-gene

topology

toestim


andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san


requires

anultram

etricgene

tree

andrecent


phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum



nthepresence

ofahigh


numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim


ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto


site

noahmreidhom


Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Guillo

trsquosUnified

Mod

elmdashpo

pulatio


Thisapproach

provides

astatistical

model

that

allowsoneto

analyzegenetic

and

phenotyp

ticdata

with

inaun

ified

mod

elandinferencefram

eworkandop

tionally

incorporateinform


nof

samplesA

Bayesianclustering

algorithm

assumes

that

each

clusterin

ageographical

domaincanbe

approxim

ated

bypo

lygons

that

arecentered

around

pointsgeneratedby

aPo

issonprocessGuillo

trsquos

Unified

Mod

elisflexiblein

term

sof

thegenetic

data

that

itcanutilize

andcapableof

accurately

delim

iting

species

Limita

tions

Genetic

andphenotypic

data


historiesfor


divergence

forneutralgenetic

markers

andadaptatio

nfora

morphological

structure

Source

Availableas

anextensionof

theRGENELAND

package(G

uillo

tet

al20

05

2012)(http

www2im

mdtudk

gigu

Geneland)

Genotyp

icandno

n-genetic

(eg

phenotypicalecolog

icalgeograph

ical

behavioral)data

STRUCTUREmdash

populatio

nassignment


Amodel-based

clustering


genotype

data

toinferpopulatio

nstructureandassign

individualsto

populatio

nsIndividualsin


probabilistically

topopulatio

nso

rjointly

totwoor

morepopulatio

nsiftheirgenotypes

indicatethatthey

areadmixedT

hemodeldoes

notassum

eaparticular

mutationprocess

anditcanbe

appliedto

mosto

fthecommonly

used

genetic

markersp

rovidedthat

they

areno

tcloselylin

ked

The


lyaccurate

assignmentsusing

modestn

umbersof

loci(Pritchard

etal2

000)T


fpopulatio

nstructurecanbe

inferred

byassessinglik

elihoodscores

orthead

hocΔKstatistic

(Evann

oet

al20

05)

Limita

tions

Identifying


genetic

clusters

ischallenging

clusters

produced

byST

RUCTUREcanbe

strongly

influenced

byvariationin

sample

sizeclusters

createdby

STRUCTUREmay

notbe

consistent

with

theevolutionary

historyof

thepopulatio


elylong

divergence

times

with

inevolutionary

lineagesTem

poraldivergence

andrelatio

nships

amongputativ

egroups


atedEquivalence

togenetic

clusters


isuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source


ldescribedin

Falush

etal(200

3)andPritchard

etal(200

0)

Genotyp

icdata

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim


ergdisequ

ilibrium

foragiven


everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition


ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram


Huelsenbeck

etal(201

1)

Genotyp

icdata


distances

ldquoAutom

aticBarcode

Gap



thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto


should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model


aYulemod


which

isthen

fitto

aninferred

sing

le-gene

topology

toestim


andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san


requires

anultram

etricgene

tree

andrecent


phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum



nthepresence

ofahigh


numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim


ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto


site

noahmreidhom


Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

Structuram

amdashpo

pulatio

nassignmentusing

Bayesianclustering

Implem

entstheclustering

algorithm

used

inST

RUCTURE(see

above)

that

clusters

samples

into

populatio

nsby

minim


ergdisequ

ilibrium

foragiven


everStructuram

aalso

includes

theadditio

nof

reversible-jum

pMCMCto

identifytheop

timalpartition


ofgenetic

datacanbe

inpu

tinto

Structuram

aandtheprog

ram

canassign

individu

alsto

populatio

nwith

orwith

outtheadmixture

Limita

tions

Tem

poraldivergence

andrelatio

nships

amongputativ

egroups

isnot

explicitlyestim

atedE

quivalence

togenetic

clusters

tospecies-levelg

roupsisuncertain

andvalid

ationapproaches

canbe

used

toassess

evolutionary

independence

ofclusters

Source

httpctegberkeleyedu

structuram


Huelsenbeck

etal(201

1)

Genotyp

icdata


distances

ldquoAutom

aticBarcode

Gap



thebarcodegap

The

methoduses

arecursiveapproach

topartition

thedata

andtestfor

sign

ificant

gapsA

BGDisfastsim

plemethodto

split

asequence

alignm

entd

atasetinto


should

becomplem

entedwith

otherevidence

inan

integrative

taxono

mic

approach

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source

httpwwwabisnv

jussieufrpublicabg

ddescribedin

Puillandreet

al(201

2)

Single-locus

sequence

alignm

ent

GMYCampbG

MYCmdashgene

tree

The

GMYC

approach

combinesacoalescent

model


aYulemod


which

isthen

fitto

aninferred

sing

le-gene

topology

toestim


andastatistical

measure

ofconfi

denceforthe

inferred

boundariesA

san


requires

anultram

etricgene

tree

andrecent


phylogenetic

relatio

nships

and

parametersof

theGMYCmod

elT

heGMYCisgenerally

stable

across

awiderangeof

circum



nthepresence

ofahigh


numbers

ofsampled

speciesandgaps

inintraspecific

sampling

theaccuracy

oftheGMYCismostsign

ificantly

affected

bythemean

populatio

nsize

relativ

eto

divergence

times

betweenthem

Limita

tions

The

GMYCmay

delim


ofhaplotypes

asindependent

lineagesandas

such

may

beproneto

over-delim

itatio

nSingle-locus

data

aloneshould

only

beused

toprov

ideaprelim

inaryperspectiveof

speciesbo

undaries

andno

tas

the

sole

evidence

inspeciescircum

scriptions

Source

(http

r-forger-projecto


site

noahmreidhom


Monaghanet

al(200

9)Fu

jisaw

aandBarrac-

loug

h(201

3)andPo

nset

al(200

6)

Sing

le-locus

ultram

etricgene

tree (con

tinued)

20 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

PTPamp

bPTPmdash

gene

tree

The

Poissontree

processes(PTP)

modelcanbe

used

toinferputativ

especiesboundaries


inputtreePT

Pcaninferputativ

especiesboundaries

that

are

consistent

with

thePS

CAnim

portantadvantageof

thismethodisthat

itmod

els

speciatio

nin

term

sof

thenumberof


circum



processof

generatin

gultram

etrictreeswhich

are

required

asan

inpu

tfor

GMYCmod

el(see

abov

e)F


Pmodel

may

outperform


methods

whenevolutionary

distancesaresm

all

Limita

tions

Single-locus

data

aloneshould

only

beused

toprovideaprelim

inary

perspectiveof

speciesboundaries

andnotas

thesole

evidence

inspecies

circum

scriptions

Source


iswebsoftwarePT


Zhang

etal(201

3)

Sing

le-locus

gene

tree

GeneiousSp

eciesDelim

itatio

nplug

-inmdash

gene

tree

Aplug-into


anexploratorytool

allowingtheuser

toassess

phylog

enetic

supportanddiagno

sabilityof

speciesdefinedas

user-selected

collections

oftaxa

onuser-sup

pliedtreesThe

plug

-incompu


gto

the

probability

oftheob

served

mon

ophyly

orexclusivity

having

occurred

bychance

ina

coalescent

processandassesses

thewith


distancesto

infertheprobability

with

which

mem

bers

ofaputativ

especiesmight

beidentifi

edsuccessfully

with

tree-based

methods

Limita

tions

The

plug-insummarizes

measuresof

phylogenetic

supportand

diagnosabilityof

speciesdefinedas


oftaxabutitdoes

not

providedefinitiv



Source

Implem

entedas

aplug-into


Mastersetal

(201

1)

Sing

le-locus

gene

tree

BPamp

Pmdashmultispecies

coalescent

model

for

speciesvalid

ation

Thisapproach

tospeciesdelim

itatio

nuses


togeneratethe



accountof

uncertaintiesdueto

unknow

ngene


processThe

methodrelieson

auser-

specified

guidetreeimplem

entin

gareversible-jum

pMarkovchainMonte

Carlo

search

ofparameter

spacethatincludes

θpopulatio

ndivergenceand

estim

ated

distributio

nsof

gene

treesfrom

multip

leloci

Limita

tions

Misspecificatio

nsof

priorsandthegu

idetree

canresultin

inflatedspeciatio


norecombinatio

nandcomputatio

nallim

itatio

nsrestrict

itsutility

with


100s

ofloci

Source


YangandRannala

(201

0)andRannala

andYang(200

3)

Multilocus

sequence

alignm

entsandgroup

mem

bership

(con

tinued)


Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

DISSE

CTmdashmultispecies

coalescent

model

forspeciesdelim

itatio

nDISSE

CTexplores

thefullspaceof

possible

clusteringsof


topologies

inaBayesianfram

ework

Toavoidtheneed

forreversible-jum

pMCMCit

uses

anapproxim

ationin

theform


nof



plem

entedas

partof

BEAST

andrequires

only

afewchanges

from

astandard

BEAST

analysisA

nalysesof

simulated

andem

piricald

atasuggestthat

themethodisshow

nto

beinsensitive

tothedegree

ofapproxim

ation

butq

uitesensitive

tootherparameters

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nItappearsthat

alargenu

mberof

sequencesarerequ

ired

todraw

firm

conclusion

sSo

urce


cmcamphttpwwwin

driid

com

dissectinbeast

htmldescribedin

JonesandOxelm

an(201

4)

Multilocus

sequence

data

SpeD


coalescent

mod

elfordiscov

ery

valid

ation

andjoint

estim

ation

Thismaxim

umlik

elihoo

dandorinform

ationtheory-based

methodwas

developedto

test

speciesboundaries

inasystem

with

existin

gsubspecies


ey20


ofthegene


forall

hierarchical

perm

utations

oflin

eage

grou

pingSp

eciesbo

undaries

arecomparedusing

Akaikeinform

ationcriteria

andphylogenetic

uncertaintyin

thespeciestree

topologies

does

notaffect

speciesdelim

itatio

nsLimita

tions


quality

ofthegene

tree

estim

ates

Source

(http


describedin

Ence

andCarstens(201

1)

Multilocus

sequence

alignm

entsandgroup

mem

bership

Browniemdash

multispecies

coalescent

model

forspeciesdelim

itatio

nThe

nonp

aram

etricheuristic

speciesdelim

itatio

napproach

implem

entedin

theprog

ram

Brownie(O

rsquoMeara

2010)jointly

sortsanonym

oussamples

into

speciesandinfers

aspeciestree

from

inputg

enetreesfrom


neven


gene

treeswill

bemoreconsistent

with

each

than

the

divergenceswith

inspecies

Limita

tions

Findingboth

theoptim

umspeciestree


remains

compu

tatio

nally

challenging

andBrowniehasbeen

show

nto

frequently

yieldincorrect

resultsT

heaccuracy

ofthemethodislik

elycorrelated

with

nodalsupportvalues

inthe

individu

algene

topo

logies

Source

httpwwwbrianom

earain

fobrownie

describedin

OrsquoM

eara

(201

0)

Individual

gene

trees

(con

tinued)

22 SD Leavitt et al

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

Table

21

(con

tinued)

Metho

dDescriptio

nInpu

tdata

BFD

mdashmultispecies

coalescent

model

for

speciesdelim

itatio

nThe

recently

developedmethod

Bayes

factor

delim

itatio

n(with

genomicdataB

FD)

combinesady

namic


forestim


bypasses

thecompu

tatio

nally


ergene

treesto

prov

idearigorous


itatio


PdataCom

petin

gspeciesdelim

itatio



this

approach

isrobustto

samplesizes(iefew

loci

andlim

itedsamples

perspecies)

and

misspecificatio

nof

thepriorforpo

pulatio

nsize

(θ)

Limita

tions

Recently


theoretical

andem

pirical

evaluatio

nSo

urce


describedin

Leacheacuteet

al(201

4)

Genom

e-wideSN

Pdata







24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References






24 SD Leavitt et al















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References















26 SD Leavitt et al












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References












28 SD Leavitt et al














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References














30 SD Leavitt et al













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References













32 SD Leavitt et al













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References













34 SD Leavitt et al






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References






References






































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

































36 SD Leavitt et al


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References


























































38 SD Leavitt et al






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References






















































40 SD Leavitt et al

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References

























































42 SD Leavitt et al








































44 SD Leavitt et al


Abstract










Acknowledgments

References








































44 SD Leavitt et al


Abstract










Acknowledgments

References

the dynamic discipline of species 2 delimitation: progress ... · paramount for conservation, human...

Documents