mitogenomic phylogeny, diversification, and biogeography

HAL Id hal-03079822httpshalumontpellierfrhal-03079822

Submitted on 14 Jun 2021

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents whether they are pub-lished or not The documents may come fromteaching and research institutions in France orabroad or from public or private research centers

Lrsquoarchive ouverte pluridisciplinaire HAL estdestineacutee au deacutepocirct et agrave la diffusion de documentsscientifiques de niveau recherche publieacutes ou noneacutemanant des eacutetablissements drsquoenseignement et derecherche franccedilais ou eacutetrangers des laboratoirespublics ou priveacutes

Distributed under a Creative Commons Attribution| 40 International License

Mitogenomic phylogeny diversification andbiogeography of South American spiny rats

Pierre-Henri Fabre Nathan Upham Louise Emmons Fabienne Justy YuriLeite Ana Carolina Loss Ludovic Orlando Marie-Ka Tilak Bruce Patterson

Emmanuel Douzery

To cite this versionPierre-Henri Fabre Nathan Upham Louise Emmons Fabienne Justy Yuri Leite et al Mitogenomicphylogeny diversification and biogeography of South American spiny rats Molecular Biology andEvolution Oxford University Press (OUP) 2017 34 (3) pp613-633 101093molbevmsw261 hal-03079822

Mitogenomic Phylogeny Diversification and Biogeography ofSouth American Spiny Rats

Pierre-Henri Fabre12 Nathan S Upham34 Louise H Emmons2 Fabienne Justy1 Yuri LR Leite5 AnaCarolina Loss5 Ludovic Orlando67 Marie-Ka Tilak1 Bruce D Patterson4 and Emmanuel JP Douzery1

1Institut des Sciences de lrsquoEvolution (ISEM UMR 5554 CNRS-IRD-UM) Universite de Montpellier Montpellier France2National Museum of Natural History Smithsonian Institution Washington DC3Ecology and Evolutionary Biology Yale University New Haven CT4Integrative Research Center Field Museum of Natural History Chicago IL5Departamento de Ciencias Biologicas Universidade Federal do Espırito Santo Vitoria Espırito Santo Brazil6Centre for GeoGenetics Natural History Museum of Denmark University of Copenhagen Copenhagen Denmark7Laboratoire drsquoAnthropobiologie Moleculaire et drsquoImagerie de Synthese Universite de Toulouse University Paul Sabatier ToulouseFrance

Corresponding author E-mail phfmouradegmailcom

Associate editor Emma Teeling

Abstract

Echimyidae is one of the most speciose and ecologically diverse rodent families in the world occupying a wide range ofhabitats in the Neotropics However a resolved phylogeny at the genus-level is still lacking for these 22 genera of SouthAmerican spiny rats including the coypu (Myocastorinae) and 5 genera of West Indian hutias (Capromyidae) relativesHere we used Illumina shotgun sequencing to assemble 38 new complete mitogenomes establishing Echimyidae andCapromyidae as the first major rodent families to be completely sequenced at the genus-level for their mitochondrialDNA Combining mitogenomes and nuclear exons we inferred a robust phylogenetic framework that reveals severalnewly supported nodes as well as the tempo of the higher level diversification of these rodents Incorporating the fullgeneric diversity of extant echimyids leads us to propose a new higher level classification of two subfamiliesEuryzygomatomyinae and Echimyinae Of note the enigmatic Carterodon displays fast-evolving mitochondrial andnuclear sequences with a long branch that destabilizes the deepest divergences of the echimyid tree thereby challengingthe sister-group relationship between Capromyidae and Euryzygomatomyinae Biogeographical analyses involving higherlevel taxa show that several vicariant and dispersal events impacted the evolutionary history of echimyids The diver-sification history of Echimyidae seems to have been influenced by two major historical factors namely (1) recurrentconnections between Atlantic and Amazonian Forests and (2) the Northern uplift of the Andes

Key words biogeography diversification Echimyidae mitogenomics Neotropics nuclear DNA fast-evolving geneCarterodon

Introduction

Thanks to a unique geological history and ecology theNeotropical biota together composes the richest realm ofbiodiversity in the world The formation of complex riverinesystems and the uplift of the Andean Cordillera has togetherwith climatic changes driven several spectacular biologicalradiations (Hoorn et al 2010 Patterson and Costa 2012)Based on geology and zoogeography (Morrone 2014) thetropical Americas can be subdivided into six major biomes(fig 1) Amazon Basin Central Andes and Guyana Shield(AM) the trans-Andean region plus Choco (CH) theNorthern Andes and northern coastal dry forests inVenezuela and Colombia (NA) Caatinga Chaco Cerradoand the Pantanal (CC) Atlantic Forest (AF) and the insularWest Indies (WI) These biomes are characterized by severallarge forested regions (Amazon Basin Atlantic Forest ChocoOrinoco watersheds and West Indies) open habitats (CC

CerradothornCaatinga and Chaco NA Pantanalthorn Llanos)and the Andean mountain range (CentralthornNorthernAndes) The major barriers currently separating the forestedbiomes were formed following climate aridification andormountain uplift during the Miocene (Hoorn et al 2010)and have substantially contributed to biotic divergence andendemism in the Neotropics The Amazon Basin forms thelargest Neotropical biome and is surrounded to the west bythe Choco wet forest (across the Northern Andes) both theOrinoco watershed and the northern coastal dry forests inVenezuela and Colombia and the coastal Atlantic Forest ofBrazil and Paraguay Disjunct distributions of terrestrial ver-tebrate species among the Amazon Basin and the AtlanticForests reflect a common geoclimatic divergence structuringin South American biotas (da Silva and Patton 1993 Pattonet al 2000 Batalha-Filho et al 2013) Biotas of those tworainforests are among the most diverse globally and are sep-arated by a diagonal region of xeric habitats comprising the

Article

The Author 2016 Published by Oxford University Press on behalf of the Society for Molecular Biology and EvolutionAll rights reserved For permissions please e-mail journalspermissionsoupcom

Mol Biol Evol 34(3)613ndash633 doi101093molbevmsw261 Advance Access publication December 26 2016 613

Dow

nloaded from httpsacadem

icoupcomm

bearticle3436132739699 by guest on 14 June 2021

Cerrado and Caatinga biomes (AbrsquoSaber 1977) Geologicaland paleontological data (Hoorn et al 2010) suggest thatareas of the Amazon Basin and Atlantic Forest have under-gone successive separations and connections since theMiocene including (1) expansion of foreland hydrologicbasins of the West Amazon basin (23ndash7 Ma see Lundberget al 1998 Wesselingh and Salo 2006 Wesselingh et al 2006)(2) potential watershed shifts between the Parana Sea and theAmazon River (12ndash7 Ma Hoorn 1996 Reuroaseuroanen and Linna1996 Lundberg et al 1998 Roddaz et al 2006) and (3)more recent Plio-Pleistocene climatic events (Costa 2003)Xeric habitats appear to have represented more labile barriersto gene flow than previously hypothesized (Patton et al 2000)as demonstrated by the relatively young phylogeographicstructure in birds and mammals of the northernNeotropics (Batalha-Filho et al 2013 Nascimento et al2013) North of the Amazon Basin the uplift of theNorthern Andes and emergence of the Orinoco watershedhas also shaped patterns of biodiversity both as a vicariantbarrier between Mesoamerica and coastal areas north of theAmazon Basin and as a heterogeneous region within whichspecies can be isolated diverge and later disperse to otherregions (ie a ldquospecies pumprdquo) (Cracraft and Prum 1988 Ron2000 Patterson and Costa 2012)

In this paleo-geographical context South American spinyrats (Mammalia Rodentia Echimyidae) have radiated acrossmultiple biomes and encompass today a vast array of lifehistories and ecomorphological adaptations includingcapacities for semi-fossorial arboreal and semi-aquatic

lifestyles (see review by Emmons et al 2015)Understanding the process of higher level diversification ofEchimyidae is at the core of several major biogeographicalquestions Historical biogeography has been posited (Leiteand Patton 2002 Galewski et al 2005) as a main driver ofechimyid rodent divergence with two major regions involvedin alternating vicariance and dispersal through time (1) east-ern South America including the Atlantic Forest and theopen formations of the Cerrado and Caatinga and (2) theAmazon Basin including the cis-Andean highland regionsand the forests of the Southern Orinoco Basin and GuianaShield The lowland rainforest of the Amazon Basin supportsmost of the extant echimyid diversity with eight arborealgenera (Dactylomys Echimys Isothrix Makalata MesomysLonchothrix Toromys Pattonomys) and the speciose terres-trial genus Proechimys (at least 22 species Emmons et al2015) The Atlantic Forest biome also harbors a diverse as-semblage of echimyid species grouped into three arborealgenera (Kannabateomys Phyllomys Callistomys) one terres-trial genus (Trinomys) and one semi-fossorial genus(Euryzygomatomys) In between the Amazon Basin andAtlantic Forest echimyids also diversified in (1) grasslandsas semi-fossorial specialists in the Cerrado and Caatinga bio-mes (Carterodon and Clyomys) and (2) within open vegeta-tion of the Caatinga Cerrado Chaco and Pantanal with theground-dwelling terrestrial genus Thrichomys (Borodin et al2006 DrsquoElıa and Myers 2014) This pattern suggests the influ-ence of vicariant processes between the Amazonian andAtlantic forests (Galewski et al 2005 Fabre Galewski et al

FIG 1 Distribution map for all genera of Echimyidae and correspondence between their ranges and the biogeographical provinces of theNeotropics

Fabre et al doi101093molbevmsw261 MBE

614

Dow


icoupcomm


2013) as well as the influence of past Andean uplift(Patterson and Velazco 2008 Upham et al 2013) offeringtwo important drivers for the genus-level diversification ofechimyid species

Previous phylogenetic results suggest the initial diversifica-tion of Echimyidae during the Early Miocene (Galewski et al2005 Upham and Patterson 2012 Fabre Galewski et al2013) including independent originations of arboreality inthe Amazon and presumably again in the Atlantic Foreston the lineage leading to Callistomys (Loss et al 2014) Themain arboreal clade of echimyids which includes taxa in bothAmazonian and Atlantic rainforest areas indicates that mul-tiple transitions have been made among these regions(Upham and Patterson 2012 Fabre et al 2013) If these obli-gate arboreal taxa could not have crossed the dry biomes ofCaatinga and Cerrado without a forest corridor thenMiocene introgressions of freshwater (eg Pebas or AcreSystems) could have played a role in the arboreal diversificat-ion of mammals as has been suggested for Chaetomys (Vilelaet al 2009) and other vertebrates (Martini et al 2007Pellegrino et al 2011 Fouquet et al 2012) As this pattern isnot well documented in mammalian lineages (but see Costa2003 and Costa and Leite 2012) additional comparisons withphylogenetic biogeographical and geological results will offera better understanding of isolation processes in this areathrough the Miocene

The role of the Northern Andes as a barrier between thelowland rainforests of the Choco and Orinoco watershedsand those of the Amazon Basin are overlooked in all currentbiogeographical models Since the late Miocene Andean or-ogeny has been an active force in major faunal separationbetween these areas (Cracraft and Prum 1988 Ron 2000Delsuc et al 2004 Hoorn et al 2010 Patterson and Costa2012 Patterson et al 2012) This mountain uplift has led to(1) cis- (eastern slope) and trans-Andean (eastern-to-westernslope) vicariant separations of several mammalian groups(Ron 2000 Patterson and Costa 2012) and (2) south-to-north divergences between the Amazon Basin and Orinocowatershed At least five echimyid genera are now endemic tothe Northern Andes including Olallamys Hoplomys andDiplomys in trans-Andean Choco rainforest and CentralAmerica and Pattonomys and Santamartamys in cis-Andean north of the Cassiquiare canals (we excludeAmazonian Pattonomys occasius a species pending revisionEmmons LH and Fabre P-H personal communication) Thetrans-Andean regions include the basins between the separ-ate branches of the northern Andes the lowland Pacificcoastal region the very wet Choco forest and the extensionof that biome into southern Central Americamdashthe barriersbetween which may have encouraged alternating events ofvicariance and dispersal for echimyids

The picture of phylogenetic ecomorphological and bio-geographical evolution of echimyids is complicated byCapromyidae the family comprising several extinct and ex-tant genera of hutias endemically distributed in the WestIndies (Davalos and Turvey 2012) Capromyids also displayvarious life history traits and adaptations including terrestrialscansorial and arboreal taxa Because capromyids have been

considered as a distinct family closely related to echimyids(Woods and Kilpatrick 2005 Upham and Patterson 2012) oreven nested within a paraphyletic echimyid assemblage(Galewski et al 2005 Upham and Patterson 2012 2015Fabre Galewski et al 2013 Fabre et al 2014) taxa fromboth families should be sampled to understand the evolutionof these taxa

Beyond biogeographical patterns the Echimyidae havebeen the focus of recent taxonomic debates (see Carvalhoand Salles 2004 Emmons 2005 Patton et al 2015 for anoverview) Pioneering taxonomies were primarily based onexternal characters and teeth occlusal patterns that arenow considered homoplasic (Candela and Rasia 2012 Fabreet al 2013) Recent molecular studies (Lara et al 1996 Leiteand Patton 2002 Galewski et al 2005 Fabre et al 2013Upham et al 2013) have indicated that the terrestrial spinyrats Proechimys and Trinomys long considered congenericbelong to different echimyid clades and are not closelyrelated despite overall morphological similarities Withinthe main arboreal clade there is a lack of phylogenetic signalat deeper nodes so that many basal relationships remainunresolved (Galewski et al 2005 Fabre Galewski et al2013 Upham et al 2013) The sampling of new molecularmarkers as well as hitherto unsequenced genera (CarterodonCallistomys Diplomys Pattonomys Santamartamys) there-fore promises to enhance our understanding of echimyidsystematics and biogeographic history (Upham andPatterson 2015)

The rise of next-generation DNA sequencing (NGS) meth-ods (eg Illumina Meyer et al 2008) has made access tomitochondrial (mito) genomes and nuclear genes from freshtissues readily available and in an increasingly time- and cost-effective manner for museum skins (Rowe et al 2011Guschanski et al 2013) High-throughput sequencing meth-odologies have for example increased the number of availablecomplete mitogenomes of mammals (Horner et al 2007Horn et al 2011 Fabre Joslashnsson et al 2013) including tworecent studies on Caviomorpha the lineage of SouthAmerican rodents to which echimyids belong (Tomascoet al 2011 Vilela et al 2013) Moreover the availability ofprobabilistic mixture models of sequence evolution (Lartillot2004 Lartillot et al 2007) allow us to maximally extract phylo-genetic information from mitochondrial sequences nucleargenes and a combination thereof

Here we used an Illumina shotgun sequencing approach(Tilak et al 2015) to obtain 38 new mitogenomes ofEchimyidae (including members of the Capromyidae andMyocastorinae) and up to five slower-evolving nuclear exonsfrom key taxa Our approach has for example generated thefirst complete mitogenomes from rare museum specimens ofendemic species in the genera Santamartamys Olallamysand Mesocapromys After combining these newly sequencedmitogenomes and nuclear markers with publically availablesequences we used model-based approaches (maximum like-lihood Bayesian inference) to infer the phylogeny ofEchimyidaethornCapromyidae and to estimate a timeframefor their diversification Then we explored their Neogene evo-lutionary history addressing the following questions

Mitogenomic Phylogeny Diversification and Biogeography of South American Spiny Rats doi101093molbevmsw261 MBE

615

Dow


icoupcomm


(1) Does the addition of new mitogenomic and nucleargene data plus enhanced sampling of arboreal taxa help toimprove the node support of the higher level echimyid andcapromyid phylogeny (2) What has been the timeframe ofthe EchimyidaethornCapromyidae diversification (3) To whatextent are geo-climatic historical events reflected in the high-er level composition of echimyid communities withinNeotropical biomes

Materials and Methods

Taxonomic SamplingWe sampled all 26 extant genera (Emmons 2005 Woods andKilpatrick 2005) in the clade composed of Echimyidae (2323including Myocastorinae) and Capromyidae (55) Biologicalsamples were obtained of both museum and fresh tissuecollections from the American Museum of Natural History(AMNH New York NY) the Field Museum of Natural History(FMNH Chicago IL) the Museum of Comparative Zoology(MCZ Cambridge MA) the Museum of Vertebrate ZoologyUniversity of California (MVZ Berkeley CA) the US NationalMuseum of Natural History (USNM-SmithsonianWashington DC) the Nationaal Natuurhistorisch Museum(Naturalis Leiden The Netherlands) the Royal OntarioMuseum (ROM Toronto Canada) the UniversidadeFederal do Espirito SantomdashAnimal Tissue Collection (UFES-CTA Vitoria Brazil) and the University of Montpellier (UM-ISEM Montpellier FR) (see supplementary table S1Supplementary Material online) In total we obtained com-plete mitogenomes for 38 taxa spanning all extant membersof Echimyidae and Capromyidae In addition we searchedNCBI (see also supplementary table S1 SupplementaryMaterial online) for mitogenomic and nuclear sequencesdocumented by voucher specimens We incorporated repre-sentatives of the octodontoid families Ctenomyidae andOctodontidae their closest Caviomorpha relatives (Uphamand Patterson 2015) to serve as outgroups We also includedone member of the families Caviidae Erethizontidae andChinchillidae to represent the other three superfamilies ofcaviomorphs as more distant outgroups to stabilize theOctodontoidea relationships

Sequencing and Assembly of Complete MitochondrialGenomesEleven museum skin samples were stored in Eppendorf tubesa subset of which was processed in the ldquoDegraded DNArdquofacility of the University of Montpellier France (dedicatedto processing low qualityquantity DNA tissue samples)Alcohol-preserved samples of fresh tissues were available for27 other taxa (table 1 and supplementary table S1Supplementary Material online) the DNA from these sam-ples was extracted in a dedicated room of the laboratory toreduce contamination risks An additional DNA extract ofSantamartamys rufodorsalis was included in the preparationof genomic libraries derived using methods described inUpham et al (2013) DNA was extracted using the DNeasyBlood and Tissue Kit (Qiagen) following the manufacturerrsquosinstructions with a final elution in water The 11 oldest

samples were extracted in small batches (of 3 samples max-imum) and a negative control was included in each batch tomonitor possible contamination The DNA was fragmentedby sonication using an ultrasonic cleaning unit (Elmasonic)The 30-ends of the obtained fragments were then repairedand double-strand filled before being ligated with adaptorsand tagged according to a cost-effective protocol for Illuminalibrary preparation (Meyer and Kircher 2010 Tilak et al 2015)Tagged DNA libraries were pooled and sequenced as singleend reads on Illumina HiSeq 2000 lanes at the GATCndashBiotechcompany (Konstanz Germany)

Raw 101-nt reads were imported in Geneious R6 (Kearseet al 2012) and adaptor fragments were removed by the ldquotrimendsrdquo utility Then a mapping of the reads on the phylogen-etically closest available mitochondrial genome was per-formed for each species The following mapping parameterswere used in the Geneious read mapper a minimum of 24consecutive nucleotides (nt) perfectly matching the refer-ence a maximum 5 of single nt mismatch over the readlength a minimum of 95-nt similarity in overlap region anda maximum of 3 of gaps with a maximum gap size of 3-ntIterative mapping cycles were performed in order to elongatethe sequence until recovery of the complete mitogenome Ahigh-quality consensus was generated and the circularity ofthe mitogenome was verified by the exact 100-nt sequencesuperimposition at the assembly extremities The number ofsatellite repeats in the control region remained underesti-mated as there were Illumina reads consisting entirely ofrepeats making it difficult to determine the exact copy num-ber New mitochondrial genomes are available under acces-sion numbers KU762015 and KU892752 to KU892788

Sequencing and Assembly of Nuclear MarkersTo complement the mitochondrial information with nuclearDNA markers we used the protocol of Fabre et al (2014 cfSupplementary Methods in the Data SupplementSupplementary Material online) to capture five partial nuclearexons for Carterodon and Callistomys (apoB GHR IRBP RAG1vWF) In brief nine primer sets from the literature were usedto amplify nuclear DNA baits from modern genomic extractsof Capromys pilorides Bait amplicons were sheared into300ndash500 bp fragments All purified amplicons were builtinto blunt-end libraries Captured libraries were amplified be-fore being pooled equimolarly and sequenced on one lane ofa Illumina MiSeq (50 bp) single-read run Another captureusing just the nuclear DNA baits was also sequenced on onelane of a HiSeq2000 run Newly acquired nuclear exons werecombined with previously published ones (Fabre et al 2014)With the exception of two taxa (Lonchothrix emiliae andSantamartamys rufodorsalis) we obtained nuclear gene se-quences for all examined species New nuclear exons areavailable under accessions KY303652 to KY303660

DNA and Protein SupermatricesWe compared our newly obtained mitogenomes and nu-clear data with sequences available from public databases(Lara et al 1996 Lara and Patton 2000 Leite and Patton2002 Galewski et al 2005 Patterson and Velazco 2008


616

Dow


icoupcomm


Meredith et al 2011 Fabre Galewski et al 2013 Uphamet al 2013) Complete mitogenomes and nuclear exonswere aligned with MUSCLE (Edgar 2004) as implementedin SEAVIEW v452 (Gouy et al 2010) and with MACSE version101b (Ranwez et al 2011) Ambiguous regions of the align-ments were cleaned with trimAl version 14 (Capella-Gutierrez et al 2009)

We combined the mitochondrial and nuclear sequencesinto three DNA supermatrices (1) the complete mitoge-nomes (47 taxa and 15896 sites 07 of missing characterstates ie calculated on the taxon x site basis) (2) the con-catenated nuclear exons (45 taxa and 5415 sites 32 of

missing character states) and (3) the complete mitogenomescombined with nuclear exons that maximize our charactersampling (47 taxa and 21311 sites 11 of missing characterstates) We also concatenated the 13 mitochondrial and 5nuclear protein-coding genes translated into amino acid (AA)sequences in a super-protein dataset (47 taxa and 5614 sitespartition ldquoAArdquo 15 of missing character states)

Phylogenetic AnalysesIndividual gene trees for the 13 mitochondrial protein-codinggenes and the 5 nuclear exons were built using maximumlikelihood (ML) with parameters and topologies estimated

Table 1 List of the Species Treated in This Study and GenBank Accession Numbers of Sequences

Taxa Collection Tissue_number Publication Accession mitogenome Tissues

Ctenomys rionegrensis mdash mdash Tomasco et al (2011) HM5441301 Fresh tissueOctodon degus mdash mdash Tomasco et al (2011) HM5441341 Fresh tissueSpalacopus cyanus mdash mdash Tomasco et al (2011) HM5441331 Fresh tissueTympanoctomys barrerae mdash mdash Tomasco et al (2011) HM5441321 Fresh tissueProechimys longicaudatus mdash mdash Tomasco et al (2011) HM5441281 Fresh tissueTrinomys dimidiatus mdash mdash Voloch et al (2013) JX3126941 Fresh tissueCallistomys pictus UFES-CTA RM-233 This paper KU892754 Fresh tissueCapromys pilorides MVZ Mammals MVZ-191417 This paper KU892766 Fresh tissueCarterodon sulcidens UFES-CTA LGA-735 This paper KU892752 Fresh tissueClyomys laticeps UFES-CTA MCNM-2009 This paper KU892753 Fresh tissueDactylomys dactylinus UFES-CTA UFROM-339 This paper KU762015 Fresh tissueDiplomys labilis USNM-Smithsonian USNM-335742 This paper KU892776 Dry study skinEchimys chrysurus UM-ISEM T-3835 This paper KU892781 Fresh tissueEuryzygomatomys spinosus UFES-CTA CTA-1028 (YL-53) This paper KU892755 Fresh tissueGeocapromys browni Naturalis-Leiden ZMAMAM23884 This paper KU892767 Dry study skinGeocapromys ingrahami USNM-Smithsonian USNM-395696 This paper KU892768 Dry study skinHoplomys gymnurus MVZ Mammals MVZ-225082 This paper KU892779 Fresh tissueIsothrix sinnamariensis UM-ISEM T-4377 This paper KU892785 Fresh tissueKannabateomys amblyonyx UFES-CTA MBML-3001 This paper KU892775 Fresh tissueLonchothrix emiliae FMNH FMNH 140821 This paper KU892786 Dry study skinMakalata didelphoides UM-ISEM T-5023 This paper KU892782 Fresh tissueMesocapromys melanurus MCZ-Harvard MCZ-34406 This paper KU892769 Dry study skinMesomys hispidus UM-ISEM T-6523 This paper KU892787 Fresh tissueMesomys stimulax Naturalis-Leiden RMNHMAM21728 This paper KU892788 Dry study skinMyocastor coypu UM-ISEM T-0245 This paper KU892780 Fresh tissueMysateles prehensilis gundlachi MCZ-Harvard MCZ-17090 This paper KU892770 Dry study skinOlallamys albicauda USNM-Smithsonian USNM-241343 This paper KU892774 Dry study skinPattonomys carrikeri ROM ROM-107955 This paper KU892783 Fresh tissuePhyllomys blainvillii UFES-CTA CTA-1205 This paper KU892756 Fresh tissuePhyllomys dasythrix UFES-CTA AC-628 This paper KU892757 Fresh tissuePhyllomys lundi UFES-CTA CTA-881 This paper KU892758 Fresh tissuePhyllomys mantiqueirensis UFES-CTA CTA-912 This paper KU892759 Fresh tissuePhyllomys pattoni UFES-CTA CTA-984 This paper KU892760 Fresh tissuePlagiodontia aedium Naturalis-Leiden RMNHMAM3865 This paper KU892771 Dry study skinProechimys cuvieri UM-ISEM T-5765 This paper KU892778 Fresh tissueProechimys roberti UFES-CTA CTA-1524 This paper KU892772 Fresh tissueSantamartamys rufodorsalis AMNH AMNH-34392 This paper KU892777 Dry study skinThrichomys apereoides MVZ Mammals MVZ-197573 This paper KU892773 Fresh tissueToromys grandis FMNH FMNH 92198 This paper KU892784 Dry study skinTrinomys albispinus UFES-CTA AL-3054 This paper KU892761 Fresh tissueTrinomys iheringi UFES-CTA ROD-156 This paper KU892762 Fresh tissueTrinomys paratus UFES-CTA CTA-588 This paper KU892763 Fresh tissueTrinomys setosus UFES-CTA YL-710 This paper KU892764 Fresh tissueTrinomys yonenagae UFES-CTA PEU-880027 This paper KU892765 Fresh tissue

NOTEmdashVoucher numbers of the newly sequenced taxa used in this study are given in gray shades AMNH American Museum of Natural History New York NY FMNH FieldMuseum of Natural History Chicago IL MCZ Museum of Comparative Zoology Harvard University Cambridge MA MVZ Museum of Vertebrate Zoology Berkeley CANaturalis Nationaal Natuurhistorisch Museum Leiden The Netherlands ROM Royal Ontario Museum Toronto Ontario Canada UFES-CTA Animal Tissue CollectionUniversidade Federal do Espırito Santo Brazil UM University of Montpellier collections Montpellier France USNM National Museum of Natural History SmithsonianInstitution Washington DC Most of these samples are vouchered and archived in accessible collections Additional information could be retrieved via the following databasewebpages httpvertnetorg


617

Dow


icoupcomm


under PAUP (Swofford 2003) version 4b10 A neighbor-joining (NJ) starting topology allowed to optimize the par-ameter values of the GTRthornCthorn invariable sites model of NTevolution The optimal topology was then identified after aheuristic search under tree bisectionndashreconnection (TBR)branch swapping Node bootstrap percentages were com-puted by ML after 100 replicates using previously estimatedparameters NJ starting tree and TBR branch swapping

Bayesian inferences were then conducted on all of theconcatenated datasets To account for potential heterogen-eity of substitution patterns within and between mitochon-drial and nuclear partitions we used the CAT mixture model(Lartillot 2004) implemented in PHYLOBAYES (Lartillot et al2009) version 33f Relative exchangeabilities between NTand AA were described under the general time reversible(GTR) model (Rodriguez et al 1990) To account foramong-site heterogeneity in the NT and AA substitutionrates we used a Gamma distribution with four discrete cat-egories (C4) For each dataset two Markov chain Monte Carloanalyses were run with PHYLOBAYES for 10000 cycles (ca8000000 generations) with trees sampled every five cyclesafter discarding the first 1000 as burn-in Convergence wasensured when the maximum difference in bipartition fre-quencies as estimated by the two chains was lt01 Nodesupport was estimated by posterior probabilities (PP) com-puted from samples of 9000 post-burn-in trees

To estimate the average rate of evolution among mito-chondrial and nuclear partitions at the DNA level we used abackbone topology (the highest PP topology inferred underthe CAT model for the mitochondrialthorn nuclear DNA con-catenation) PAUP (Swofford 2003) was used to estimatebranch lengths with the same backbone topology under aGTRthornCthorn invariable sites model of NT evolution across all22 partitions (13 mitochondrial protein-coding genes 2 ribo-somal RNAs the control region all tRNAs combined into asingle partition and the 5 nuclear exons) The super distancematrix (SDM) approach (Criscuolo et al 2006) was used toestimate the 22 average rates of evolution of the DNA parti-tions through the use of the resulting trees converted intoadditive distance matrices by computing the path-length be-tween each pair of taxa (see Ranwez et al 2007 for details)The resulting SDM rates were standardized into relative evo-lutionary rates (RER) with respect to the slowest-evolvingmarker set to RERfrac14 1

Molecular Dating AnalysesDivergence times were estimated from the mitochon-drialthorn nuclear NT and AA supermatrices to provide a tem-poral framework of the echimyid radiation A Bayesianrelaxed molecular clock method with rate autocorrelationalong branches (Thorne et al 1998 Thorne and Kishino2002 Lepage et al 2007) was used to estimate divergencedates while accounting for changes in evolutionary rateover time and allowing for mixture models of sequence evo-lution to account for heterogeneities in patterns of evolutionamong mitochondrial and nuclear partitions We again usedPHYLOBAYES (Lartillot et al 2009) to estimate the divergencedates within South American spiny rats using the Bayesian

consensus topology obtained from DNA data (see also fig 2)The best previously estimated Bayesian topology was used torun the molecular dating analysis under the CATthornGTRthornCmixture model for both NT and AA datasets To calibrate themolecular clock we selected five fossil constraints alreadyconsidered in previous studies on rodents (Upham et al2013 Patterson and Upham 2014) First the most recentcommon ancestor (MRCA) of CavioideandashErethizontoideawith an age range between 313 and 459 Ma The minimumbound was based on the fossil Andemys termasi (voucherSGOPV 2933) which is conservatively the oldest stemCavioidea representative and which could even be con-sidered as a stem Dasyproctidae (313ndash336 Ma) from theTinguirirican SALMA late Eocene-early Oligocene (Fields1957 Vucetich et al 1993 2010 Bertrand et al 2012 Dunnet al 2013 Verzi 2013) The older bound was assumed to beno younger than the oldest stem members of Caviomorpha(Canaanimys and Cachiyacuy Barrancan SALMA middleEocene 41 Ma [409ndash459 Ma] Antoine et al 2012)

Secondly the Octodontidae MRCA was constrained usingthe oldest crown fossil Pseudoplateomys innominatus (MACN8663) from the Huayquerian SALMA late Miocene (68ndash90 Ma) We set a soft maximum (upper 95) prior age of118 Ma for this octodontid node based on oldest occurrenceof stem octodontids such as Acarechimys from the LaventanSALMA late Miocene (Candela and Noriega 2004 Verzi 2008)

We subsequently used three ingroup calibrations We con-strained the divergence between Trinomys andClyomysthorn Euryzygomatomys to range from 60 to 118 Maas based on Theridomysops parvulus (MACN-Pv8379) theoldest stem taxon to Euryzygomatomysthorn Clyomys fromthe Huayquerian SALMA late Miocene (Hussain et al 1978Wyss et al 1993 Schultz et al 2002 Verzi 2008) We then usedthe oldest arboreal echimyid Maruchito trilofodonte (MLP 91-IV-1-22) to date the MRCA of the arboreal clade (EchimyiniFabre et al 2014) with a range of 157ndash242 Ma This fossilfrom the Colloncuran SALMA157 Ma middle Miocene isthe oldest stem taxon between Phyllomys and Echimys incladistic studies (Vucetich et al 1993 Macphee 2005 Verzi2008 Dunn et al 2013) Finally we calibrated the MRCA ofThrichomys and the clade including Hoplomys ProechimysCallistomys and Myocastor using the fossil Pampamysemmonsae (GHUNLPam 2214) from the HuayquerianSALMA late Miocene and the age range of 60ndash118 Ma(Olivares et al 2012 Verzi 2013 Verzi et al 2015)

Biogeographical AnalysesWe used LAGRANGE Cthornthorn to elucidate ancestral biogeograph-ical patterns using Dispersal Extinction CladogenesisthornX(DECX) historical biogeography models (Ree et al 2005 Reeand Smith 2008 Matzke 2014) This approach allowed us toinfer the number of dispersalvicariance events among generaof Echimyidae throughout the Neogene Probabilistic biogeo-graphic analyses using DECX (Ree et al 2005 Beeravolu andCondamine 2016) optimize reconstructed ancestral areasupon internal nodes LAGRANGE Cthornthorn calculates maximumlikelihood estimates of the ancestral states (range inheritancescenarios) at speciation events by modelling transitions


618

Dow


icoupcomm


between discrete states (biogeographic ranges) along phylo-genetic branches as a function of time In addition to DECrsquostwo free parameters d (dispersal) and e (extinction) DECX ischaracterized by two additional parameters vicariance (VI)and rapid anagenetic change (RAC) that were not consideredin the previous DEC model (Ree et al 2005) and DECthorn Jmodel (Matzke 2013) VI is a standard parameter whereasRAC contrasts the rapid cladogenetic change model imple-mented in DECthorn J The distribution of echimyid ranges fromSouth and Central America were defined based on recentrevisions available in Patton et al (2015) We subsequentlydivided regions into smaller biogeographic entities to obtainhigher resolution in the ancestral area of origin Due to ourgenus-level approach we joined together open habitat areasinto a single region as well as Amazonia and the Guiana shieldUsing tectonic inferences notably the evolution of pastAmazonian landscapes (Hoorn et al 2010) we assigned six

areas for the DEC analysis Amazon Basin Central Andes andGuiana Shield (AM) the trans-Andean region plus Choco(CH) the Northern Andes and northern coastal dry forestsin Venezuela and Colombia (NA) Caatinga Chaco Cerradoand the Pantanal (CC) Atlantic Forest (AF) and the WestIndies (WI)

We inferred echimyid areas of origin and the direction andsequence of colonization among these areas on the chrono-gram inferred from AA under the CAT mixture model Onegreat advantage of the DECX model (eg over parsimony-based models such as DIVA Ree et al 2005) is the ability tomodel transition rates and connectivity scenarios We ac-knowledge that such a constrained model might be skewedespecially if transition rates and connectivities between areasare not estimated properly In consequence two models wereconsidered in the present study First a simple (ldquouncon-strainedrdquo) model was computed under equal exchange

FIG 2 Bayesian topology obtained from the nucleotide dataset (NT) concatenation of the 13 protein-coding and the 5 nuclear genesPosterior probabilities as computed under the CATthornGTRthornC mixture model by Phylobayes are indicated on each node See figure 3 fordeeper exploration of the results The tree is rooted by Caviomorpha taxa CtenomyidaethornOctodontidae (outgroup) and Cavia porcellusCoendou sp and Chinchilla lanigera (not shown) Systematic names clade colors and letter labels for the major echimyid clades are givenalong the phylogeny Illustrations of Plagiodontia aedium by Toni Llobet from Wilson et al (2016) Illustrations of Echimyidae byFrancois Feer from Emmons and Feer (1997)


619

Dow


icoupcomm


probabilities between areas Second to test whether transi-tions among regions were correlated with major geologicalevents we used a ldquostratifiedrdquo biogeographic scenario thatallowed for (1) limited dispersal between unconnected areasand (2) connectivity changes between major geologicalevents Movement between unconnected areas is prohibitedin this model by setting the dispersal parameter d (probabilityof dispersal between nonadjacent areas) equal to 0 We setdfrac14 1 between all adjacent areas Next we stratified themodel based on the Neotropical geology data presented inHoorn et al (2010) employing five transition matrices basedon successive geological time-intervals The West Indies hasbeen separated from mainland South America during theentire Neogene (Iturralde-Vinent and Macphee 1999) Weallow connectivity between the Amazon Basin and AtlanticForest during the period between 12 and 7 Ma due to thepotential presence of savannah corridors and the Parana sea(Hoorn et al 2010) Connectivity between the Amazon Basinand Northern South America ended between 12 Ma and thepresent due to major Andean uplift events which isolatedthe trans- and cis-Andean regions from each other and theOrinoco watershed from the cis-Andean and SouthernAmazon basin fauna (Hoorn et al 2010) The widespreadNeotropical genus Proechimys was coded as AMthornNA dueto our limited species sampling for this genus Beeravolu andCondamine (2016) reported that DECX outperforms DECthorn Jin the case of stratified models The methodology of its cousinDECthorn J is usually better at dealing with nonstratified modelsHowever in our case DECX performed better than DECthorn Jboth on unconstrained and stratified models (supplementarytable S3 Supplementary Material online) With our AIC andAkaike weight results the differences between DECX andDECthorn J were not statistically different for the unconstrainedand significant under the stratified model (supplementarytable S3 Supplementary Material online) We chose not touse DECthorn J instead implementing a stratified model thattakes into account geological history and because the J par-ameter (jump dispersal) is more suited to island systems(Beeravolu and Condamine 2016)

Ecomorphological AnalysesWe mapped the ecomorphological character of locomotionlifemode onto the phylogeny for all Echimyidae genera basedon account in Patton et al (2015) Coding for locomotionlifestyle was as follows (1) semi-arboreal scansorial or arbor-eal (Callistomys Capromys Dactylomys Diplomys EchimysIsothrix Kannabateomys Lonchothrix MesocapromysMakalata Pattonomys Plagiodontia Mesomys MysatelesOlallamys Phyllomys Santamartamys Toromys) (2) semi-fossorial (Clyomys Euryzygomatomys Carterodon) (3) terres-trial (Proechimys Geocapromys Hoplomys ThrichomysTrinomys) and (4) semi-aquatic (Myocastor) Ancestral char-acter state reconstructions were performed in the phytoolspackage in the R programming language (R Core Team 2016)using the ldquomakesimmaprdquo function (Bollback 2006 Revell2012) This stochastic mapping approach iteratively recon-structs states at all ancestral nodes in the phylogeny usingrandom draws from a k-state Markov model (MkthornC) as

conditioned on the best-fitting model of transition ratesamong states and integrates over the uncertainty in observednodal states We used maximum-likelihood (ML) to estimatethe best-fitting model with the function ldquoacerdquo in the apepackage (Paradis et al 2004) and then simulated 1000 sto-chastic maps on the maximum clade credibility tree fromPHYLOBAYES Plotting of the SIMMAP results were performedusing the function plotSimmap in R

Results and Discussion

Mitogenomes Help to Resolve a Thorny Star-PhylogenyWe screened the new DNA sequences for potentially anom-alous assemblies andor contamination by building phylogen-etic trees on individual mitochondrial genes and nuclearexons We did not detect any well-supported(1gt PPgt 095) topological conflict among the inferred mito-chondrial and nuclear gene trees (table 2 and supplementaltable S2 Supplementary Material online) For this reason andbecause both concatenation and coalescence analyses usingsophisticated models of sequence evolution and in-depthtree search routines generally yield concordant phylogenetichypotheses for mammals (Gatesy et al 2016) we conductedall subsequent inferences on mitochondrial andor nuclearsupermatrices and we did not use coalescent-based investi-gation We relied on the combination of faster-evolving mito-chondrial sequences for reconstructing relationships amongspecies and genera with slower-evolving nuclear exons fordeeper nodes of the echimyid and capromyid tree We there-fore approximated the species phylogeny through the com-bination of gene alignments Thanks to cumulative signalpotentially contributed by all individual genes the superma-trix approach here appeared to be the best way to providesupport for the trickiest nodes of the spiny rat phylogeny Thisdistinction is emphasized because higher level phylogeneticrelationships within Echimyidae have been consistently enig-matic using both single and combined mitochondrial andnuclear markers but never evaluated using mitogenomesand all extant genera (Lara et al 1996 Leite and Patton2002 Galewski et al 2005 Fabre et al 2013 Upham et al2013)

With more than 21000 unambiguously aligned mitochon-drial plus nuclear positions our analysis provides improvedsupport for the phylogenetic position of several echimyidgenera First we confirmed that Myocastor belongs withinthe Echimyidae (contra Woods and Kilpatrick 2005)Combined analysis of all markers show Myocastor nestedwithin Echimyinae a subfamily comprised of an arboreal tribe(Echimyini) and a terrestrial tribe (Myocastorini) (fig 2) Thiswell-supported Myocastorini tribe unexpectedly includesCallistomys as sister taxa to Myocastor (Loss et al 2014)Both taxa are the sister group to a moderately supportedgroup containing Thrichomys and Hoplomysthorn Proechimys(Node R) This position for Myocastor is compatible withthe classification of Echimyidae including Myocastor (sensuWoods et al 1992 Galewski et al 2005) giving further supportto the McKenna and Bell (1997) who recognize


620

Dow


icoupcomm


Myocastorinae within Echimyidae We confirm three well-supported major clades (1) the Capromyidae consisting ofWest Indian endemics and closely related to Echimyidae(Fabre et al 2014) (2) the East Brazilian old endemics herereferred to as Euryzygomatomyinae and (3) the Echimyini(ie all arboreal echimyids except Callistomys) branchingwith the tribe Myocastorini We confirm that Trinomys isplaced within Euryzygomatomyinae closely related to thefossorial Clyomys and Euryzygomatomys genera (Lara andPatton 2000) Thus despite its prior synonymy withProechimys over most of the 20th century Trinomys is onlydistantly related to the Myocastorini contradicting previousclassifications

The arboreal Echimyini clade (node E) is always recoveredand we propose the best-supported hypothesis to date for thegenus-level branching patterns within this radiation We re-covered several previously suggested subclades (1) the bam-boo rats (Dactylomys Olallamys and Kannabateomys nodeO) (2) the Northern Andean endemics Santamartamys and

Diplomys (Diplomys also occurs in trans-Andean areas) nodeN) (3) the Amazonian cis-Andean clade Toromys andPattonomys (node J) (4) the three remaining Echimyinaetaxa (Echimys Makalata Phyllomys node K) and theMesomysLonchothrix clade (Node G) Our comparative ana-lysis of mitogenomes provides strong statistical support for (1)the sister relationship between Dactylomys and Olallamys(node P) to the exclusion of Kannabateomys (2) a sister rela-tionship between Phyllomys and Echimys (node L) and (3) theassociation of Nodes J and K into a subclade that includes themost diverse arboreal genera (Makalata and Phyllomys) (NodeI) followed by their branching with bamboo ratsthornDiplomysand Santamartamys (node H) Within this Echimyini clade wealso found weak supportmdashthough consistently recovered bydifferent markersmdashfor the association of the isolatedlineage Isothrix with LonchothrixthornMesomys (node F butsee also fig 3)

Regarding the gene sampling our expanded set of molecularmarkers appeared to be central to reconstructing several as-pects of Echimyidae and Capromyidae phylogeny (fig 3)Individual mitochondrial genes such as the widely used cyto-chrome b marker in mammals or the cytochrome oxidase Ibarcode can show limited phylogenetic resolution particularlyat deeper nodes While several mitochondrial genes can beconcatenated to form a ldquosupergenerdquo the minimal number ofgenes necessary to significantly resolve the phylogenetic struc-ture remains contentious with an average of three to fivemitochondrial genes seeming sufficient to reach a resolutionsimilar to that derived from the analysis of the complete mito-genome (Duchene et al 2011 Havird and Santos 2014)Because the selection of genes serving as a proxy for the mito-genomic tree can vary with the particular clades of interest ouruse of complete mitochondrial genomes for the echimyids andcapromyids allowed us to assemble comparable data acrosstaxonomic groups while avoiding a highly ldquogappyrdquo supermatrixof characters In the present study the combination of com-plete mitogenomes with nuclear exons yields a 21311-sitesupermatrix with only 11 taxon-by-site missing data

The estimation of the relative evolutionary rates amongthe 22 mitochondrial and nuclear partitions (table 3) indi-cated that (1) the slowest-evolving markers are the nuclearexons apoB and RAG1 followed by the 2-fold faster-evolvingGHR IRBP and vWF (RERfrac14 20ndash23) (2) the mitochondrialtRNAs and rRNAs evolved 5ndash6 times faster (3) the 13 mito-chondrial protein-coding genes evolved ca 22ndash45 timesfaster (RERfrac14 223 228 424 and 446 respectively foratp8 cytb cox1 and nd4l) and (4) the control region evolved373-fold faster (see Supplementary Material online) Theseestimates show that our mitochondrial and nuclear DNAsupermatrix of ca 21000 unambiguously aligned sites con-tains partitions with contrasting evolutionary dynamicsThis combination of slower- and faster-evolving markerstherefore provided phylogenetic resolving power at the spe-cies level by the faster-evolving sequences (eg withinPhyllomys and Trinomys or among closely related generaof the Node E) as well as at deeper taxonomic levels by theslower-evolving genes (eg Capromyidae Echimyinae andEuryzygomatomyinae)

Table 2 Phylogenetic Support Values of the Major Clades Identifiedfrom the Mitochondrial andor Nuclear Supermatrix Analyses UnderProbabilistic Approach for Nucleotides

Nodes PP Congruence Genes

Mitonuc Mito Nuclear

A 100 100 100 5 vs 0 V A I G RB 099 095 029 1 vs 4 v a i G rC 100 100 100 3 vs 1 V _ I G rD 100 100 100 3 vs 2 V A I g rE 100 100 099 3 vs 2 v A I G rF 094 098 010 3 vs 0 v _ _ g rG 100 100 H 100 100 077 2 vs 1 v _ I G rI 100 100 100 3 vs 0 V _ _ G RJ 100 100 100 2 vs 0 V A _ _ _K 100 100 074 3 vs 2 V A I g rL 100 100 089 3 vs 2 V A I g rM 099 098 096 1 vs 1 v _ _ G _N 100 100 O 100 100 099 1 vs 1 V _ _ g _P 100 100 048 1 vs 1 v _ _ G _Q 100 100 100 5 vs 0 V A I G RR 086 045 008 2 vs 2 v _ I G rS 100 100 100 4 vs 0 V _ I G RT 100 100 076 4 vs 1 V A i G RU 091 069 010 1 vs 2 v _ _ G rV 100 100 100 5 vs 0 V A I G RW 100 100 100 3 vs 1 v A _ G RX 100 100 100 3 vs 2 v a I G RY 100 100 046 1 vs 2 v _ I _ r

NOTEmdashNodes are labelled with letters (see figs 2 and 3)PP posterior probabilities computed under the CAT mixture model by PhylobayesMitonuc mitonuclear DNA supermatrix Mito mitogenome DNA supermatrixNuclear nuclear DNA supermatrixThe column ldquocongruencerdquo indicates the number of nuclear gene trees that respect-ively recover versus do not recover the corresponding clade of the mito-nuclearsupermatrix tree Gene names V von Willebrand Factor exon 28 (vWF) A apoli-poprotein B exon 26 (APOB) I Interphotoreceptor retinoid-binding protein exon 1(RBP3) G Growth hormone receptor exon 10 (GHR) and R Recombination acti-vating protein 1 gene (RAG1) The upper case letter indicates congruence with theCATthornGTRthornC (Phylobayes) mito-nuclear topology while the lower case indicatesa different topology A question mark indicates that genes are absent for the taxaunder focus precluding the recognition of the corresponding clade


621

Dow


icoupcomm


Table 3 DNA Substitution Rates among the 22 Mitochondrial and Nuclear Partitions

Genome Markers Sites Cumul sites SDM rates Scaled rates Remarks

mtDNA 12S 880 880 123 61 mdashmtDNA 16S 1406 2286 124 62 mdashmtDNA ND1 960 3246 654 325 mdashmtDNA ND2 1041 4287 584 290 mdashmtDNA CO1 1545 5832 852 424 mdashmtDNA CO2 684 6516 668 332 mdashmtDNA AT8 192 6708 448 223 mdashmtDNA AT6 681 7389 635 316 31 nt AT8-AT6 overlapmtDNA CO3 783 8172 676 336 1 nt overlap AT6-CO3mtDNA ND3 345 8517 704 350 mdashmtDNA ND4L 297 8814 896 446 7 nt overlap NDL-ND4mtDNA ND4 1380 10194 469 233 mdashmtDNA ND5 1812 12006 574 286 mdashmtDNA ND6 534 12540 720 358 4 nt overlap ND5-ND6mtDNA CYB 1140 13680 459 228 mdashmtDNA CR 765 14445 749 373 mdashmtDNA tRNA1OL 1494 15939 103 51 mdashnucDNA apoB 1173 17112 020 10 mdashnucDNA GHR 806 17918 039 20 mdashnucDNA RAG1 1071 18989 020 10 mdashnucDNA IRBP 1207 20196 039 20 mdashnucDNA vWF 1153 21349 046 23 mdash

NOTEmdashGenome Mitochondrial (mt) or nuclear (nuc) Markers Name of the partition (CR control region tRNAthornOL 22 tRNAs and the L-strand origin of replication) SitesNumber of sites of the corresponding alignment Cumul sites Cumulative number of sites in the DNA supermatrix SDM rates Relative substitution rates among markers asmeasured by SDM Scaled rates Rates scaled with respect to the slowest partition (apob set to 10)

FIG 3 Bayesian topologies obtained from the nucleotide (NT) and amino acid (AA) concatenations of the 13 protein-coding and the 5 nucleargenes Results without and with Carterodon are highlighted herein Posterior probabilities as computed under the CATthornGTRthornC mixture modelby Phylobayes are indicated on each node See figure 2 for nucleotide results including Carterodon The tree is rooted by caviomorph taxa that werepruned from the trees in order to produce these figures Taxonomic names clade colors and letter labels for the major echimyid clades are givenalong the phylogeny U0 represents the alternative relationships for the node U Capromyidaethorn Euryzygomatomyinae


622

Dow


icoupcomm


Regarding the sampling of DNA markers we built trees fromthe mitochondrial-only or nuclear-only supermatrix under theCATthornGTRthornC model The matrilineal (mitochondrial DNA)topology produced the same four major clades Capromyidae(PPfrac14 100) Euryzygomatomyinae (PPfrac14 095) Myocastorini(PPfrac14 100) and Echimyini (PPfrac14 100) with the latter twostrongly grouped together (PPfrac14 100) and comprisingEchimyinae Deeper relationships are not resolved with allEchimyidae except Carterodon branching together(PPfrac14 064) as the sister-group of all Capromyidae plusCarterodon (PPfrac14 069) The biparental (nuclear DNA) top-ology also evidenced major clades Capromyidae (PPfrac14 100)Euryzygomatomyinaethorn Carterodon (PPfrac14 087) Myocastorini(PPfrac14 100) and Echimyini (PPfrac14 099) with the latter twostrongly grouped together (PPfrac14 100) Deeper relationshipsare not resolved with a trifurcation involving CapromyidaeEuryzygomatomyinaethorn Carterodon and MyocastorinithornEchimyini The mitogenomic tree and the nuclear exon treeare congruent apart from the phylogenetic position ofCarterodon (but see ldquoDiscussionrdquo below)

Regarding the molecular character sampling we cannotrule out the possibility that mitogenomic sites reached sub-stitution saturation for resolving the echimyid phylogeny be-cause of the fast evolutionary rates of the mitochondrialprotein-coding sequences However when conducting thesame inferences at the AA level we recovered a topologyvirtually identical to the one reconstructed from the NTsupermatrix suggesting that substitution saturation had little(if any) impact on our findings (see figs 2 or 3) The onlytopological difference at the protein level is the modificationof the basal trifurcation of Echimyidae with CapromyidaethornCarterodon branching as the sister group of all other generaalthough with a low PP value Other differences involveweaker PP supports at the protein level for the few nodesthat were not fully supported at the DNA level (eg the PPvalue decreases for node R) This might reflect the lowerpercentage of variable sites available for the probabilistic in-ference in the AA versus NT supermatrices at the DNA level491 and 381 of the sites are respectively variable andparsimony-informative whereas these proportions decreaseto 420 and 275 at the protein level

Regarding taxon sampling we sampled all of the recog-nized genera of Echimyidae and Capromyidae and analysedthese in the most rigorous molecular phylogenetic frameworkto date (but see also Upham and Patterson 2015) Samplingall major lineages with at least one species per genus enablesus to better understand the evolution of the spiny rat familyby minimizing the number of isolatedmdashand potentiallylongermdashbranches in the resulting phylogeny This phylogenytherefore benefits from an improved detection of multiplesubstitution events leading to a more accurate estimation ofboth topology and branch lengths which could otherwisehinder the delineation of evolutionary signal versus noise(Poe 2003) Our sampling reduced this systematic error inorder to produce a more stable phylogenetic picture withexhaustive generic coverage (Delsuc et al 2005)

Regarding the inference model an inherent difficulty inmitochondrial phylogenetic analyses results from differing

substitution patterns due to different functions of DNA (pro-tein-coding ribosomal RNA tRNA and noncoding) as well asdifferent NT and AA compositions (eg 12 H-strand protein-coding genes versus the single L-strand ND6) This difficulty ishere compounded by the further concatenation of geneticallyunlinked nuclear exons that also display different NT or AAcompositions substitution patterns and rates of replace-ment Mixture models like the CAT model implemented inPHYLOBAYES circumvent these problems by dynamically recon-structing independent NT profiles that individually fit subsetsof similarly evolving sites (Lartillot et al 2007) Moreover theyavoid the need to remove supermatrix partitions that wouldotherwise violate model assumptions We also note that thetimetree estimation relies heavily upon accurate branchlength measurement under a specified model of sequenceevolution Distribution by the CAT mixture model of thealignment columns into different categories accounts forsite-specific NT and AA preferences (Lartillot 2004) Branchlengths estimated under the CAT model will therefore be lessaffected by saturation because of a better ability to detectmultiple hits and will integrate the substitution pattern het-erogeneity exhibited by mitochondrial and nuclear loci(Lartillot et al 2007)

As a result of the improved sampling of mitochondrial andnuclear markers molecular characters and taxa and throughthe use of the CAT mixture model most nodes of theEchimyidaethornCapromyidae phylogeny are now strongly sup-ported Exceptions are four nodes that are unresolved despitethe use of ca 21000 sites (figs 2 and 3) Genera involved inthese weaker supported nodes are Carterodon (node U)Thrichomys (node R) Isothrix (node F) and to a lesser extentthe Diplomysthorn Santamartamysthorn the three bamboo ratgenera (node M) Three genera are represented in thephylogeny as isolated branches (Carterodon Thrichomysand Isothrix) A fourth branchmdashthe one subtendingDiplomysthorn Santamartamysmdashalso behaves similarly as itssubdivision is not very deep the branch lengths from nodeN to terminal taxa are of roughly similar length to the internalone between nodes M and N (figs 2 and 3) Phylogeneticplacement of isolated branches might be difficult but wouldbenefit from further improvements to taxon sampling Futurestudies might therefore incorporate a richer species samplingwithin Thrichomys (five species are recognized) and Isothrix(six species are recognized) but to our knowledge the branchleading to Diplomys labilis could only be subdivided by addingD caniceps whereas Santamartamys rufodorsalis is mono-typic It is noteworthy that the analysis of Upham (2014)and Upham and Patterson (2015) based on ca 5000 nucleo-tides but with enhanced taxon sampling (eg all 6 Isothrix)recovered stronger support for some of these same groupings(nodes M and R) but not others (nodes F and U)

The placement of Carterodon is perhaps the most prob-lematic as this monotypic genus exhibits the longest branchin the analysis of complete mitogenomes and the apoB GHRIRBP RAG1 and vWF exons In the tree inferred from themito-nuclear DNA supermatrix (cf figs 2 and 3) the mediandistance from the EchimyidaethornCapromyidae MRCA to thetips is 089 NT substitution per site whereas it is ca twice as


623

Dow


icoupcomm


long for Carterodon (161) At the AA level a similar contrastis observed with a MRCA-to-tips median distance of 028 AAsubstitution per site versus 049 for Carterodon Moreoverboth analyses of either the matrilineal or the biparentalsupermatrices again yield trees with the longest branch lead-ing to Carterodon Regarding the phylogenetic position ofCarterodon we caution that saturation could hamper its ac-curate placement and possibly impact the branching patternwith its closest relatives We note that the 13 mitochondrialand the five nuclear protein-coding sequences sampled forCarterodon do not display premature stop codons in theopen reading frames suggesting that pseudogenes were notincorporated in the alignments and are not involved in thelong-branch attribute of this taxon However we can postu-late that several nonexclusive adaptive and neutral proc-esses may have increased the evolutionary rate of themitochondrial and nuclear markers here under focus Forexample animal mitochondrial genomes have been shownto be recurrently impacted by episodes of positive selection(James et al 2016) and by life-history traits driven muta-tional biases (Jobson et al 2010) Positive selection neutralprocesses or a combination thereof may therefore lead tothe inference of long branches in the phylogenetic treeInvestigations at the population level would allow thesehypotheses to be explored but the task would be compli-cated by the need for samples from several individuals ofthe rare Carterodon sulcidens

Topologies reconstructed from all mitochondrial plusnuclear characters agree to link Carterodon withCapromyidae (PPfrac14 091 and 085 with NT and AA respect-ively) whereas the relationships among the three clades ofEchimyidaethornCapromyidae are either unresolved (trifurca-tion with NT cf fig 2) or receive moderate support for anassociation of Euryzygomatomyinae with Echimyinae to theexclusion of CarterodonthornCapromyidae After excludingCarterodon from additional inferences we observed thatthe phylogenetic position of Capromyidae shifted to abranching with Euryzygomatomyinae with increased sup-port (PPfrac14 088 with NT and 093 with AA fig 3) whereasMyocastorini remained tightly linked to Echimyini(PPfrac14 100 and 099 respectively) The same trend is alsoobserved when Carterodon is removed from either the matri-lineal-only or the biparental-only analyses Capromyidaecluster with Euryzygomatomyinae (PPfrac14 085 andPPfrac14 062 respectively) Clearly without Carterodon theresolution of the phylogeny is improved in a consistentmanner and even provides increased support comparedwith previous studies (however PPlt 095 is still viewed cau-tiously) Since the studies of Woods and Howland (1979)and Woods et al (2001) hypotheses of evolutionary affin-ities between hutias and the ancestral stock of spiny ratshave been debated on molecular grounds Galewski et al(2005) analysed four markers (three mitochondrial onenuclear) and found weak signal for a branching of their sin-gle capromyid Capromys with Clyomysthorn Euryzygomatomysand Trinomys (ML bootstrapfrac14 47) Upham and Patterson(2012) analysed four markers (one mitochondrial three nu-clear) and found the sister relationship of Capromys to all

Echimyidae but again with low support (ML boot-strapfrac14 43) Fabre et al (2013) analysed eight markers(three mitochondrial five nuclear) and found increased sup-port for Capromysthorn Euryzygomatomyinae (ML boot-strapfrac14 72) albeit with reduced taxon sampling in thishutiathorn East Brazilian clade Fabre et al (2014) used an ex-panded taxon sampling within hutias with the same eightmarkers and found increased support for CapromyidaethornEuryzygomatomyinae (ML bootstrapfrac14 76 and PPfrac14 100)Upham and Patterson (2015) analysed five markers (twomitochondrial three nuclear) for all genera but again foundan unresolved relationship among CapromyidaeEuryzygomatomyinae and Carterodon (PPfrac14 028 ML boot-strapfrac14 41 for a CarterodonthornCapromyidae relationship)

In the present study and for the first time analysis ofthe complete mitogenomes together with five nuclearmarkers for all genera of hutias and spiny ratsmdashbutexcluding Carterodonmdashsupports the above-mentionedCapromyidaethorn Euryzygomatomyinae relationshipsHowever including the long Carterodon branch destabilizedthe topology (fig 3) and prevents any taxonomic conclu-sions from being made We suspect that the deepest rela-tionships among Echimyidae and Capromyidae areimpacted by long-branch attraction phenomena involvingthe long branch of Carterodon as well as the long branches ofthe close CtenomyidaethornOctodontidae outgroups andorthe more distant Caviidaethorn Erethizontidae outgroupsWhen Carterodon is incorporated in the mitochon-drialthorn nuclear DNA analyses Echimyidae would thereforebe paraphyletic unless the hutias are included as a subfamilywithin Echimyidae an arrangement previously advocated byEllerman (1940) We hope that future work including add-itional nuclear and slowly evolving markers will offer gain amore favourable signalnoise ratio profiting from the virtuesof a phylogenomic approach (Philippe et al 2005)Additionally efforts to sequence DNA from recently extinctand allied subfossil taxa (egBoromys Brotomys) could provefruitful for parsing the long branch of Carterodon For nowand until definitive phylogenetic evidence to the contrarycan be ascertained we conservatively recommend maintain-ing the family-level status of hutias within Capromyidae

Transitions between the Atlantic Forest and AmazonBasin ldquoBusyrdquo Corridors in the MioceneOur molecular dating and biogeographic analyses (table 4 andfig 4) suggest that geographic opportunities constitute theprimary driver of echimyid higher level divergences with themost likely biogeographic reconstruction including at least 1event of vicariance and 14 dispersals explaining the currentdistribution of genera (DECX lnLfrac14622) Our biogeo-graphic inferences reveal transitions among easternNeotropical regions of Atlantic Forest Caatinga andCerrado throughout the Neogene (fig 4) plus AtlanticForestndashAmazon Basin splits among echimyid lineages (nodesO L T) from 46 to 113 Ma (meanfrac14 76 Ma) This scenariosuggests that rainforest corridors would have recurrently ap-peared throughout the Miocene to allow for dispersals andorvicariances events leading to genus-level divergences


624

Dow


icoupcomm


CaatingathornCerrado and the Atlantic Forest were involvedsequentially as a corridor andor a barrier to dispersal Thisarea of high endemism includes several divergent echimyidlineages from both open (eg Carterodon) and closed habitats(eg Trinomys Callistomys) that are not found elsewhere inthe Neotropics Echimyids in both Atlantic Forest andCaatingathornCerrado have a higher frequency of endemismthan those in the Amazon Basin northern Neotropics andWest Indies Relative to the Amazon the Atlantic Forest hastwo arboreal lineages that evolved arboreality independently(Callistomys and the Echimyini)

The open-country corridors of Caatinga and Cerradomeanwhile have triggered uncommon echimyid specializa-tions to fossoriality but they have also preserved ancientldquosurvivingrdquo forest and savanna lineages found nowhere elsein the Amazon Basin source pool (eg Carterodon and someTrinomys) With only two inferred Miocene dispersals (NodesO and U) Caatinga and Cerrado open habitats have also beeninvolved in two genus-level divergences of echimyids(Nascimento et al 2013) These historical events co-occurred with locomotory shifts from terrestriality to (1)fossoriality (Nodes C and V see Clyomys Euryzygomatomysand Carterodon) and (2) rock and bush-climbing capacities(Node R Thrichomys) The majority of echimyid species areforest adapted and some competition with their octodon-toid sister group (families Ctenomyidae and Octodontidaewhich are mostly open-country and montane forms in south-ern South America) has been suggested to possibly explain

this differential success between open and closed habitats(Upham and Patterson 2012 Nascimento et al 2013Upham 2014 Arnal and Vucetich 2015)

Geographic isolation extinctions and habitat fragmenta-tion have shaped the modern diversity of the Atlantic Forest(Costa 2003 Batalha-Filho et al 2013 Leite et al 2016) Theevolutionary relationships among Atlantic Forest echimyidgenera are particularly scattered across their family topology(fig 4) The composition of the echimyid communities in theCaatinga Cerrado and Atlantic Forest as well as their deepdivergence within the topology provide evidence of ancientand recurrent links among these biomes Based on the scarcepalynological record the Caatinga and Cerrado biomes origi-nated in the early or mid-Miocene (Pascual and Jaureguizar1990 Hoorn et al 2010) and have acted ever since as a semi-permeable barrier between the two major rainforest biomesThis poorly documented botanical transition (Colinvaux andDe Oliveira 2001) occurs at the end of the middle Mioceneclimatic optimum and the start of the ensuing cooling event(Zachos et al 2001) and is then followed by major changes inthe placental mammal fauna (Flynn 1998 Croft et al 2009Antoine et al 2016) The Caatinga and Cerrado dry corridormight thus be seen as an eastndashwest transition zone forNeotropical biota However the fauna of the AtlanticForest biome is closely tied to Caatinga and Cerrado asevidenced by some members of the Atlantic Forest terres-trial spiny rats (Trinomys) which also occur in the Caatingaand Cerrado (ie T albispinus and T yonenagae Attias et al2009 Pessoa et al 2015) Our results converged withQuaternary fossil record from Lagoa Santa in MinasGerais that documented faunal shifts between Cerradoand Atlantic Forest faunas (ref documented by PeterLund Pardi~nas et al 2016) Small mammals from thesecaves document Callistomys as well as several sigmodon-tines such as Blarinomys Brucepattersonius and Castoriaboth Atlantic forest endemic and Pseudoryzomys chiefly aCerrado endemic (Voss and Myers 1991 Emmons andVucetich 1998 Pardi~nas et al 2016) As Thrichomys andCarterodon sister-relationships are poorly supported (figs2 and 3) potential links between these open habitat taxaand other echimyid from Amazon Basin and or AtlanticForest cannot be confirmed for these lineages

These eastern Neotropical biomes have an old isolationhistory compared with Northern South America with spor-adic episodes of connections The pathway to and fromAmazon and Atlantic Forests are usually divided into (1)an ancient pathway on the southern margin of Cerradoand (2) a young Plio-Pleistocene pathway through NEBrazil (Caatinga and Cerrado) (Costa et al 2000 Batalha-Filho et al 2013 Leite et al 2016) Different events might beresponsible for these faunal exchanges within these regionsFirstly the edge of the Brazilian Shield might have been con-nected throughout the Miocene from the northern Chaco inMato Grosso (Brazil) to western Bolivia and Paraguay andfrom Rondonia southeast to Parana (Brazil) Also the axialgroove of the western Amazon Basin caused repeated inlandfreshwater and marine expansions (Webb 1995 Hoorn et al2010) Throughout the Miocene marine incursions occurred

Table 4 Molecular Dating Results from the ConcatenatedMitochondrialthornNuclear Dataset for Both Nucleotides (NT) andAmino Acids (AA)

Nodes Nucleotides (NT) Amino acids (AA)

A 136 [157ndash121] 152 [171ndash136]B 127 [147ndash117] 113 [139ndash116]C 48 [59ndash39] 48 [63ndash36]D 133 [139ndash109] 144 [151ndash123]E 120 [126ndash98] 137 [131ndash103]F 104 [121ndash91] 120 [134ndash106]G 6 3[76ndash51] 73 [89ndash58]H 98 [103ndash87] 116 [131ndash103]I 79 [92ndash68] 92 [107ndash79]J 53 [65ndash44] 71 [86ndash57]K 62 [75ndash52] 71 [85ndash58]L 56 [67ndash46] 64 [79ndash52]M 94 [108ndash83] 111 [127ndash98]N 33 [41ndash26] 45 [57ndash35]O 66 [78ndash55] 78 [92ndash65]P 56 [67ndash47] 62 [76ndash49]Q 97 [113ndash87] 113 [122ndash102]R 94 [109ndash83] 109 [118ndash98]S 55 [67ndash46] 74 [86ndash61]T 88 [103ndash78] 102 [113ndash88]U 127 [1473ndash114] 144 [164ndash126]V 76 [91ndash63] 88 [106ndash71]W 35 [46ndash27] 45 [63ndash32]X 21 [29ndash15] 29 [42ndash20]Y 14 [20ndash10] 29 [35ndash16]

NOTEmdashLetters refer to the nodes in figure 1 The mean age of each node is given inmillion years ago (Ma) together with the lower and upper bounds of the 95credibility intervals derived from the Bayesian relaxed molecular clock analysisThe fossil constraints are indicated in the Material and Methods


625

Dow


icoupcomm


both in the south along the Parana basin (Cozzuol 1996) andin the north along the Amazon palaeobasin (Reuroaseuroanen et al1995 Hoorn 1996) These recurrent shifts in the watershed ofAmazonia as well as the emergence of inland seas might have

connected or isolated Amazonian and Atlantic rainforest viadispersal corridors such as flooded savannah and rainforests(Roddaz et al 2006 Wesselingh and Salo 2006 Wesselinghet al 2006) Any of these complex water ingressions could

FIG 4 Ancestral-area reconstructions of Neotropical Echimyidae The tree is a chronogram (relaxed molecular clock) based on a PHYLOBAYES

Markov Chain Monte Carlo (MCMC) analysis of the combined data set (see also table 4) The rodent outgroups are not shown (full tree provided assupplementary figures S1 and S2 Supplementary Material online) Maximum-likelihood ancestral area reconstructions were conducted from thischronogram The result from our DECX stratified and constrained model is presented here (see ldquoMaterials and Methodsrdquo for details) Theemployed geographical distribution for each taxon is presented to the right of the taxon names Colored squares indicate the geographical originof a given node Bidirectional arrows indicate vicariance and single arrows dispersal The map indicates the inferred dispersal and colonizationroutes of members of the echimyids as well as the different biogeographical hypotheses of our DECX model The branch colors indicate the resultsof the stochastic mapping results for the life-mode characters (see legends and ldquoMaterials and Methodsrdquo but see also supplementary figure S3Supplementary Material online) Taxonomic names and lettered labels for the major echimyid clades are given along the phylogeny


626

Dow


icoupcomm


have driven the generic splits of arboreal taxa CallistomysKannabateomys Phyllomys and the isolation of terrestrialnode C (Trinomys)

A Recent Vicariant Barrier to the North The LateMiocene Andean WallConstituting a long and high physical barrier separating trop-ical lowland biotas the Northern Andean uplift has been themost recent orogenic event in the Neotropics (Hoorn et al2010) This major event resulted in the separation of echi-myids from trans- (West) and cis- (East) Andean Amazonianregion as well as from the Carribean coastal region Most ofthe trans-Andean echimyid genera are arboreal and have anorthern Amazonian genus counterpart thus highlightingtheir common evolution related to the uplift of the Andes(eg Olallamys and Dactylomys Node O) With some fewhighland exceptions (Upham et al 2013 Patton et al 2015)arboreal echimyids are lowland specialists that are more spe-ciose and abundant at lower elevations Due to their limiteddispersal abilities across highland and unforested habitatsarboreal echimyids constitute potential candidates forAmazonian rainforest vicariance similar to some other pla-cental mammals (Cortes-Ortiz et al 2003 Delsuc et al 2004Patterson and Costa 2012 Gibb et al 2015) The NorthernAndean uplift has mainly triggered the provincialism ofCarribean coastal areas north of the Amazon Basinthorn trans-Andean among arboreal lineages as revealed by sister-relationships between arboreal genera (figs 2 and 4) Weidentified only one genus-level divergence among terrestriallineages Hoplomys and Proechimys (cf Node S Myocastoriniclade) compared with three splits within the main arborealclade (Node E) If the Andes have generated such genericdiversity it might also be a diversity pump in the easternpart of cis-Andean region at shallower taxonomic levels(Upham et al 2013) but that idea awaits further studies ata finer geographic scale involving of Dactylomys MakalataProechimys and Isothrix from the same region Several species(eg Isothrix barbarabrownae and Dactylomys peruanus)occur at altitudesgt2000 m in mossy or montane rainforestin this area where elevation and isolation associated withmontane habitat heterogeneity have been hypothesized todrive divergence and speciation (Bates and Zink 1994Patterson and Costa 2012 Upham et al 2013)

Our analysis confirms an Amazon Basin origin for most ofthese genera with several cismdashtrans AndeanthornNorthernAndes pairs with Choco HoplomysmdashAmazonianProechimys trans Andean Diplomysthorn northern AndesSantamartamys clademdashAmazonian ancestors of node Mincluding northern Andes OlallamysmdashAmazonianDactylomys and Northern Carribean coastal PattonomysmdashAmazonian Toromys We already noted the exception of theWest Amazonian Pattonomys occasius which represents alineage closely related to East Amazonian Toromys(Emmons LH and Fabre P-H personal communication)Northern Neotropical echimyid assemblages are more re-cently derived from their Amazon Basin ancestors comparedwith taxa inhabiting eastern biomes The strength of theAmazon riverine barrier as well as greater proximity and

similarity of rainforest habitats might explain this closefaunistic pattern compared with the older and more phylo-genetically clustered eastern biomes This endemism is alsorecovered and paralleled in bats (Koopman 1982) andsigmodontine rodents (Weksler 2009 Parada et al 2013Salazar-Bravo et al 2013)

From the Late Oligocene northern South America hasexperienced major tectonic events related to the Andes upliftincluding the first (Late Oligocene and Early Miocene) andsecond (Late Miocene) Bolivian crises as well as some recentPliocene uplifts (Sempere et al 1990 Lovejoy et al 2006Hoorn et al 2010) North Andean echimyid splits (nodes JP N S) range from 86 to 26 Ma (meanfrac14 56 Ma) which is arather short time frame This brief time window and multiplegeneric splits likely reflect a significant impact of the secondBolivian crisis as a major driver of vicariance All these generamight have been widespread prior to the start of NorthernAndean orogeny Emergence of the Northern Andes and re-sulting trans-Andean isolation might have triggered the di-vergence of some of the rarest montane endemics(Santamartamys rufodorsalis and Olallamys edax) fromevents like the uplift of the Sierra Nevada de Santa Marta(8 Ma Hoorn et al 2010) but a recent study pointed out anOligocene age with some recent uneroded Pleistocene uplifts(Villagomez et al 2011) The rise of these montane rainforestsbetween the Magdalena and Pacific Slope ridges along withthe isolation of the arid Maracaibo and Orinoco Basins mighthave also played a major role on species level diversification ofechimyids west of the Andes and in isolated northern coastalregions Looking at species diversity patterns the followinggenera have been isolated since the second Bolivian crisis andspeciated into three areas of endemism the Choco andCentral American forest to the west (Diplomys Olallamysalbicaudus Proechimys and Hoplomys) and the NorthernAndean and Carribean coastal area (Olallamys edaxSantamartamys Proechimys Pattonomys) The highly diver-gent genera of Proechimys Makalata and Mesomys representtarget model organisms for a future look at recent dispersalalong the Northern Andes and to test for the role of morerecent Andean orogenic events

Conclusion Echimyid Higher Level DiversificationBiogeographic Events and Adaptive DivergenceThe family Echimyidae is considered as a model biologicalradiation given the rapid and early diversification of spinyrat lineages and their dietary and life-mode diversity (Leiteand Patton 2002 Patton et al 2015) From our ancestral statereconstruction (fig 4 branch colours) only one ecologicalshift from terrestrial to arboreal lifestyles (Node E) is involvedin the early diversification of major lineages A second morerecent acquisition of the arboreal lifestyle is observed forCallistomys an event that coincides with a simultaneous shiftinto semi-aquatic adaptation in the southern Neotropics forits sister genus Myocastor (Emmons and Vucetich 1998)Flooded systems might have favoured such a divergence be-tween semi-aquatic (Myocastor) and arboreal (Callistomys)habits Arboreal adaptations have been hypothesized to co-incide with the first occurrence of flooded forests during the


627

Dow


icoupcomm


Miocene likely due to the presence of inland flooded habitatsduring episodes of the Pebas or Acre Systems or even theParana sea (Galewski et al 2005 Hoorn et al 2010) Despite itsmorphological distinctiveness (see Emmons and Vucetich1998) the arboreal soft-furred Callistomys was previously al-lied with other spiny tree rats in Echimyini (Carvalho andSalles 2004 Candela and Rasia 2012 Verzi 2013 Verzi et al2015) They also share high-crowned hypselodont molarssuggesting a shared herbivorous ancestry (Emmons andVucetich 1998) However our mitogenomic analysis confirmsthe result of Loss et al (2014) that Callistomys lies outside thearboreal clade and also strongly supports a sister-group rela-tionship of Callistomys with Myocastor Both taxa join theMyocastorini clade together with terrestrial ProechimysHoplomys and Thrichomys Moreover the monophyly ofCallistomysthornMyocastor is reinforced by a synapomorphicfeature of their chromosomes as both genera share the2nfrac14 42 karyotype (Gonzalez and Brum-Zorrilla 1995Ventura et al 2008) Evident homoplasies of dental occlusalpatterns hampered previous attempts to recover affinitiesamong arboreal echimyids (Carvalho and Salles 2004Candela and Rasia 2012) Further analyses taking into accountour new phylogenetic background will likely help clarifyhypotheses of primary homology among cheekteeth charac-ters across the diversity of extant and extinct echimyid ro-dents For example entirely laminar molar occlusal patternsappear to have arisen independently in Phyllomys andDiplomys lineages and of the latter Santamartamys aloneshares joined wishbone-shaped lophs with the sister-groupof dactylomyines (Node O) suggesting that it retains a moreplesiomorphic occlusal pattern (Emmons 2005)

Historical opportunities from geographic vicariance anddispersal appear to play a primary role in the diversificationamong echimyid genera perhaps to a greater extent thantheir subsequent acquisition of locomotory adaptationsWe inferred that several echimyid ancestors from theAmazon Basin Central Andean region and Guyana Shieldexperienced either repeated vicariant events or independent

dispersals between Atlantic Forest Cerrado Caatinga and theOrinoco watershed and trans-Andean region (fig 4) Thispattern is well illustrated by the bamboo rat generaDactylomys Kannabateomys and Olallamys which occur ineach of the tropical rainforest areas (figs 2 and 4) Ecologicaland morphological convergences also appear to be involvedas illustrated by the aforementioned dual acquisition of ar-boreality and some remarkably convergent cheekteeth pat-terns in Diplomys and Phyllomys (Emmons 2005 Candela andRasia 2012) Despite their ecological diversity of diet and loco-motion echimyids seem to be a morphologically conservativelineage Plesiomorphies in both arboreal and terrestrial taxaformerly allocated to Eumysopinae (eg Trinomys andProechimys) and even within Echimyini (eg Makalata andToromys but see Candela and Rasia 2012 Fabre et al 2013)have hampered past efforts to reconstruct the evolutionaryhistory of the family using morphology

Mirroring the ecological diversification of the Echimyidaeas a whole the mid-Miocene colonization of the West Indiestriggered an insular radiation of hutias (Fabre et al 2014)Present-day hutias do not represent their past diversity andthe fossil record harbours at least 21 extinct species belongingto extant hutia genera and 3 extinct genera (Davalos andTurvey 2012) along with other extinct rodents that may beclosely related (Boromys Brotomys and Heteropsomys) Theseextinct and extant capromyids display a wide range of lifehistories and ecomorphological adaptations including bothscansorial (Geocapromys) and terrestrial (Isolobodon) formsthey appear to also document independent acquisitions orloss of arboreal lifestyles (fig 4 Plagiodontia andMysatelesthornMesocapromys) This island diversity has beengreatly impacted by extinctions related to human activities(Turvey et al 2007) hindering our view of their past diversity(Davalos and Turvey 2012)

Future directions include the need to improve the species-level sampling in phylogenetic analyses of Echimyidae andCapromyidae Thanks to the availability of priceless speci-mens from natural history collections and high-throughput

Table 5 Proposed Higher Level Classifications for Echimyids Based on Our Phylogenetic Results

Classification

Hypothesis 1 Hypothesis 2 Molecular tribes Genera N species GenBank

EchimyidaeCapromyidae Capromyinae Plagiodontini Plagiodontia 1 1 (100)

Capromyini Capromys GeocapromysMesocapromys Mysateles

9 8 (88)

Incertae sedisfamily

Incertaesedis subfamily

Carterodon 1 1 (100)

Echimyidae Euryzygomatomyinae Clyomys Euryzygomatomys Trinomys 12 10 (83)Echimyinae Myocastorini Callistomys Hoplomys Myocastor

Proechimys Thrichomys30 12 (40)

Echimyini Dactylomys Diplomys EchimysIsothrix Kannabateomys LonchothrixMakalata Mesomys OlallamysPattonomys Phyllomys SantamartamysToromys

46 35 (76)

NOTEmdashN species represent the number of species described in Wilson and Reeder (2005) and Patton et al 2015 GenBank represent the number of species available in GenBankPercentage represents the proportion of species included in our study


628

Dow


icoupcomm


DNA sequencing methodologies under-sampled andor rarelineages (table 1) are likely to be sequenced soon for bothnuclear and mitochondrial genes Compared with previoushigher level classifications the initial star-phylogeny (Laraet al 1996 Leite and Patton 2002) has now been completelyscrutinized leaving few open questions at this taxonomic level(but see ldquoDiscussionrdquo) The Miocene radiation of Echimyidaeand Capromyidae is rather young (lt20 Ma) when comparedwith older splits of other ctenohystrican rodent families(Patterson and Upham 2014) but is comparable in age totheir sister clade OctodontidaethornCtenomyidae (164 Ma[137ndash189]) In view of our phylogenetic results showing shortinternal branches for the initial diversification of Echimyidaeand Capromyidae their taxonomic status at the family orsubfamily level needs further evaluation with larger sets ofnuclear DNA markers We therefore propose two alternativeclassifications (table 5) which summarize our evolutionaryhypotheses and which should provide a refined phylogeneticframework for subsequent in-depth studies of their genomicsmolecular evolution and macroevolution

Taxonomy

Euryzygomatomyinae new subfamGenus content Clyomys Euryzygomatomys TrinomysDefinition Euryzygomatinae members of which all share anorigin in the eastern part of Brazil close to the Atlantic ForestAll of them are spiny have lower deciduous premolar 4 with 5lophids 4 (Trinomys) or 3 (Clyomys Euryzygomatomys)lophids on the lower molar 1 well-connected lophs arepresent on the cheek teeth 3 molar roots anchor the uppermolars elongate lower and upper incisor roots reduced zygo-matic arch with a slightly concave dorsal margin a ventrallyexpanded jugal with much reduced scarcely salient inferiorprocess Molecular synapomorphies include the amino-acidProline at position homologous to site 294 of the human GHR(protein accession AAI364971) the amino-acid Leucine atposition homologous to site 1318 of the human VWF (pro-tein accession AAH222581)

Myocastorini new tribeGenus content Callistomys Hoplomys Myocastor

Proechimys ThrichomysDefinition Myocastorini members of which all share longupper incisor roots (except Callistomys) mid- to long-sizedlower incisor roots all of them share 4 (CallistomysThrichomys) or 5 (Hoplomys Myocastor Proechimys) lophidson the lower deciduous premolar 4 3 roots anchor the uppermolars well-connected lophs characterize the cheek teethMembers exploit a diverse array of lifestyles including terres-trial (Hoplomys Proechimys Thrichomys) arboreal(Callistomys) and amphibious (Myocastor) adaptations

Supplementary MaterialSupplementary data are available at Molecular Biology andEvolution online

AcknowledgmentsWe thank Myriam Boivin Fabien Condamine FredericDelsuc Laurent Marivaux Jim Patton Mark Springer andtwo anonymous referees for discussions andor correctionsconcerning this article We thank Francois Catzeflis for accessto biological resources and collections We are grateful to thefollowing people and institutions for granting access to studyskins and samples Paula Jenkins Samantha Oxford KatherineDixey and Roberto Portela Miguez (BMNH) Darrin LundeNicole Edmison and Kristofer Helgen (NMH SmithsonianInstitution) Eileen Westwig Neil Duncan and Robert Voss(AMNH) Chris Conroy and Jim Patton (MVZ Berkeley)Geraldine Veron Violaine Nicolas and Christiane Denis(MNHN) Steve van Der Mije (RMNH) Hans Baagoslashe andMogens Andersen (ZMUC) and Leonora Costa (UFES) Wewould like to thank Francois Feer Toni Llobet Ana ConesaElisa Badia the team of the Handbook of the Mammals of theWorld for providing us permission to use the echimyid illus-trations for figure 2 Wewould liketo thankAlexandra Bezerrato allow us to use her picture of Carterodon sulcidens in figure3 We thank Christelle Tougard for providing access toldquoPlateforme ADN degraderdquo LHE thanks the SmithsonianInstitution for continuous support of her research P-HF ac-knowledges the SYNTHESYS Foundation for funding his workin the BMNH collections (GB-TAF-2735 GB-TAF-5026 andGB-TAF-5737) P-HF was funded by a Marie-Curie fellowship(PIOF-GA-2012-330582-CANARIP-RAT) YLRL acknow-ledges continuous financial support from ConselhoNacional de Desenvolvimento Cientıfico e Tecnologico(CNPq Brazil) and Fundac~ao de Amparo a Pesquisa eInovac~ao do Espırito Santo (FAPES Brazil) NSU and BDPacknowledge a grant from the National Science Foundation(DEB-1110805) for funding part of the molecular work P-HFand ED acknowledges funding support of the French-Brazilian GUYAMAZON program and the ldquoInvestissementsdrsquoAvenirrdquo grant managed by Agence Nationale de laRecherche (CEBA ref ANR-10-LABX-25-01) This publicationis the contribution No 2016-256 of the Institut des Sciences delrsquoEvolution de Montpellier (UMR 5554-CNRS-IRD)

ReferencesAbrsquoSaber AN 1977 Os domınios morfoclimaticos da America do Sul

Primeira aproximac~ao Geomorfologia 531ndash23Antoine P-O Marivaux L Croft DA Billet G Ganeroslashd M Jaramillo C

Martin T Orliac MJ Tejada J Altamirano AJ et al 2012 MiddleEocene rodents from Peruvian Amazonia reveal the pattern andtiming of caviomorph origins and biogeography Proc R Soc LondB 2791319ndash1326

Antoine P-O Salas-Gismondi R Pujos F Ganeroslashd M Marivaux L 2016Western Amazonia as a hotspot of mammalian biodiversitythroughout the Cenozoic J Mammal Evol 1ndash13

ArnalMVucetichMG 2015Mainradiationevents inPan-Octodontoidea(Rodentia Caviomorpha) Zool J Linn Soc 175587ndash606

Attias N Raıces DSL Pessoa LM Albuquerque H Jord~ao-Nogueira TModesto TC Bergallo HG 2009 Potential distribution and new re-cords of Trinomys species (Rodentia Echimyidae) in the state of Riode Janeiro Zoologia (Curitiba) 26305ndash315

Batalha-Filho H Fjeldsa J Fabre P-H Miyaki CY 2013 Connections be-tween the Atlantic and the Amazonian forest avifaunas representdistinct historical events J Ornithol 15441ndash50


629

Dow


icoupcomm


Bates JM Zink RM 1994 Evolution into the Andes molecular evidence forspecies relationships in the genus Leptopogon The Auk 111507ndash515

Beeravolu CR Condamine F 2016 An extended maximum likelihoodinference of geographic range evolution by dispersal local extinctionand cladogenesis bioRxiv 038695

Bertrand OC Flynn JJ Croft DA Wyss AR 2012 Two new taxa(Caviomorpha Rodentia) from the early oligocene tinguiririca fauna(Chile) Am Mus Nov 37501ndash36

Bollback JP 2006 SIMMAP Stochastic character mapping of discretetraits on phylogenies BMC Bioinformatics 788

Borodin PM Barreiros-Gomez SC Zhelezova AI 2006 Reproductiveisolation due to the genetic incompatibilities between Thrichomyspachyurus and two subspecies of Thrichomys apereoides (RodentiaEchimyidae) Syst Biodivers 49159ndash167

Candela AM Noriega JI 2004 Los coipos (Rodentia CaviomorphaMyocastoridae) del ldquoMesopotamienserdquo (Mioceno tardıoFormacion Ituzaingo) de la provincia de Entre Rıos ArgentinaInsugeo Misc 1277ndash82

Candela AM Rasia LL 2012 Tooth morphology of Echimyidae(Rodentia Caviomorpha) homology assessments fossils and evolu-tion Zool J Linn Soc 164451ndash480

Capella-Gutierrez S Silla-Martınez JM Gabaldon T 2009 trimAl a toolfor automated alignment trimming in large-scale phylogenetic ana-lyses Bioinformatics 251972ndash1973

Carvalho GAS Salles LO 2004 Relationships among extant and fossilechimyids (Rodentia Hystricognathi) Zool J Linn Soc 142445ndash477

Colinvaux PA De Oliveira PE 2001 Amazon plant diversity and climatethrough the Cenozoic Palaeogeogr Palaeoclimatol Palaeoecol16651ndash63

Cortes-Ortiz L Bermingham E Rico C Rodrıguez-Luna E Sampaio IRuiz-Garcıa M 2003 Molecular systematics and biogeography ofthe Neotropical monkey genus Alouatta Mol Phylogenet Evol2664ndash81

Costa LP Leite YLR Fonseca GAB Fonseca MT 2000 Biogeography ofSouth American forest mammals endemism and diversity in theAtlantic Forest Biotropica 32872ndash881

Costa LP Leite YLR 2012 Bones clones and biomes the history andgeography of recent neotropical mammals Chicago University ofChicago Press

Costa LP 2003 The historical bridge between the Amazon and theAtlantic Forest of Brazil a study of molecular phylogeographywith small mammals J Biogeogr 3071ndash86

Cozzuol MA 1996 The record of the aquatic mammals in southernSouth America Munch Geowiss Abh 30321ndash342

Cracraft J Prum RO 1988 Patterns and processes of diversificationspeciation and historical congruence in some neotropical birdsEvolution 42603

Criscuolo A Berry V Douzery EJP Gascuel O 2006 SDM a fast distance-based approach for (super)tree building in phylogenomics FrontBiogeogr 55740ndash755

Croft DA Anaya F Auerbach D Garzione C MacFadden BJ 2009 Newdata on miocene neotropical provinciality from Cerdas BoliviaJ Mammal Evol 16175ndash198

DrsquoElıa G Myers P 2014 Sobre las Thrichomys (HysticognathiEchimyidae) paraguayas la distincion de Thrichomys fosteriThomas 1903 Therya 5153ndash166

da Silva MNF Patton JL 1993 Amazonian phylogeography mtDNAsequence variation in arboreal echimyid rodents (Caviomorpha)Mol Phylogenet Evol 2243ndash255

Davalos LM Turvey ST 2012 Bones clones and biomes the history andgeography of recent neotropical mammals Chicago University ofChicago Press

Delsuc F Brinkmann H Philippe HE 2005 Phylogenomics and the re-construction of the tree of life Nat Rev Genet 6361ndash375

Delsuc F Vizcaıno SF Douzery EJP 2004 Influence of Tertiary paleo-environmental changes on the diversification of South Americanmammals a relaxed molecular clock study within xenarthransBMC Evol Biol 411

Duchene S Archer FI Vilstrup J Caballero S Morin PA 2011Mitogenome phylogenetics the impact of using single regions andpartitioning schemes on topology substitution rate and divergencetime estimation Kolokotronis S-O editor PLoS One 6e27138

Dunn REMadden RH Kohn MJ Schmitz MD Stromberg CAE CarliniAA Re RGH Crowley J 2013 A new chronology for middle Eocenendashearly Miocene South American Land Mammal Ages Geol Soc AmBull 125539ndash555

Edgar RC 2004 MUSCLE multiple sequence alignment with high accur-acy and high throughput Nucleic Acids Res 321792ndash1797

Ellerman JR 1940 The families and genera of living rodents Vol 1 Rodentsother than Muridae London British Museum (Natural History)

Emmons LH 2005 A revision of the genera of arboreal Echimyidae(Rodentia Echimyidae Echimyinae) with descriptions of two newgenera In Mammalian Diversification From Chromosomes toPhylogeography (A Celebration of the Career of James L Patton)In Lacey E Myers P editors University of California Publications inZoology series 133 pp 247ndash310

Emmons LH Feer F (1997) Neotropical rainforest mammals a fieldguide Chicago University of Chicago Press

Emmons LH Vucetich MG 1998 The identity of Wingersquos Lasiuromysvillosus and the description of a new genus of echimyid rodent(Rodentia Echimyidae) Am Mus Nov 32231ndash12

Emmons LH Leite YLR Patton JL 2015 Family Echimyidae In Patton JLPardi~nas UFJ DrsquoElıa G editors Mammals of South America Vol 2Chicago University of Chicago Press

Fabre P-H Galewski T Tilak M Douzery EJP 2013 Diversification ofSouth American spiny rats (Echimyidae) a multigene phylogeneticapproach Zool Scr 42117ndash134

Fabre P-H Joslashnsson KAJ Douzery EJP 2013 Jumping and gliding rodentsMitogenomic affinities of Pedetidae and Anomaluridae deducedfrom an RNA-Seq approach Gene 531388ndash397

Fabre P-H Vilstrup JT Raghavan M Sarkissian Der C Willerslev EDouzery EJP Orlando L 2014 Rodents of the Caribbean originand diversification of hutias unravelled by next-generation museo-mics Biol Lett 1020140266ndash20140266

Fields RW 1957 Hystricomorph rodents from the late Miocene ofColombia South America GFF 80246

Flynn J 1998 Recent advances in South American mammalian paleon-tology Tree 13449ndash454

Fouquet A Recoder R Teixeira M Jr 2012 Molecular phylogeny andmorphometric analyses reveal deep divergence between Amazoniaand Atlantic Forest species of Dendrophryniscus Mol PhylogenetEvol 62826ndash836

Galewski T Mauffrey J-F Leite YLR Patton JL Douzery EJP 2005Ecomorphological diversification among South American spinyrats (Rodentia Echimyidae) a phylogenetic and chronological ap-proach Mol Phylogenet Evol 34601ndash615

Gatesy J Meredith RW Janecka JE Simmons MP Murphy WJ SpringerMS 2016 Resolution of a concatenationcoalescence kerfuffle par-titioned coalescence support and a robust family-level tree forMammalia Cladistics 1ndash38

Gibb GC Condamine FL Kuch M Enk J Moraes-Barros N Superina MPoinar HN Delsuc F 2015 Shotgun mitogenomics provides a refer-ence phylogenetic framework and timescale for living XenarthransMol Biol Evol 33msv250ndashmsv642

Gonzalez S Brum-Zorrilla N 1995 Karyological studies of the SouthAmerican rodent Myocastor coypus Molina 1782 (RodentiaMyocastoridae) Rev Chil Hist Nat 68215ndash226

Gouy M Guindon S Gascuel O 2010 SeaView Version 4 a multiplat-form graphical user interface for sequence alignment and phylogen-etic tree building Mol Biol Evol 27221ndash224

Guschanski K Krause J Sawyer S Valente LM Bailey S Finstermeier KSabin R Gilissen E Sonet G Nagy ZT et al 2013 Next-generationmuseomics disentangles one of the largest primate radiations FrontBiogeogr 62539ndash554

Havird JC Santos SR 2014 Performance of single and concatenated setsof mitochondrial genes at inferring metazoan relationships relativeto full mitogenome data PLoS One 9e84080


630

Dow


icoupcomm


Hoorn MC Wesselingh FP Steege ter H Bermudez MA 2010 Amazoniathrough time Andean uplift climate change landscape evolutionand biodiversity Science 330927ndash931

Hoorn MC 1996 Miocene deposits in the Amazonian foreland basinScience 273122ndash122

Horn S Durka W Wolf R Ermala A Stubbe A Stubbe M Hofreiter M2011 Mitochondrial genomes reveal slow rates of molecular evolu-tion and the timing of speciation in beavers (Castor) one of thelargest rodent species PLoS One 6e14622

Horner DS Lefkimmiatis K Reyes A Gissi C Saccone C Pesole G 2007Phylogenetic analyses of complete mitochondrial genome se-quences suggest a basal divergence of the enigmatic rodentAnomalurus BMC Evol Biol 71ndash12

Hussain ST Brujin HD Leinders JM 1978 Middle Eocene rodents fromthe Kala Chitta Range (Punjab Pakistan) P K Ned Akad Wet B81101ndash112

Iturralde-Vinent MA Macphee RDE 1999 Paleogeography of theCaribbean region implications for Cenozoic biogeography BullAm Mus Nat Hist 2381ndash95

James JE Piganeau G Eyre-Walker A 2016 The rate of adaptive evo-lution in animal mitochondria Mol Ecol 2567ndash78 doi 101111mec13475

Jobson RW Dehne-Garcia A Galtier N 2010 Apparent longevity-relatedadaptation of mitochondrial amino acid content is due to nucleo-tide compositional shifts Mitochondrion 10540ndash547

Kearse M Moir R Wilson A Stones-Havas S Cheung M Sturrock S BuxtonS Cooper A Markowitz S Duran C et al 2012 Geneious Basic Anintegrated and extendable desktop software platform for the organ-ization and analysis of sequence data Bioinformatics 281647ndash1649

Koopman KF 1982 Biogeography of the bats of South America ThePymatuning Symposia in Ecology Special Publication SeriesPymatuning Laboratory of Ecology University of Pittsburgh 61ndash539

Lara MC Patton JL da Silva MNF 1996 The simultaneous diversificationof South American echimyid rodents (Hystricognathi) based oncomplete cytochrome b sequences Mol Phylogenet Evol 5403ndash413

Lara MC Patton JL 2000 Evolutionary diversification of spiny rats (genusTrinomys Rodentia Echimyidae) in the Atlantic Forest of Brazil ZoolJ Linn Soc 130661ndash686

Lartillot N 2004 A Bayesian mixture model for across-site heterogene-ities in the amino-cid replacement process Mol Biol Evol211095ndash1109

Lartillot N Brinkmann H Philippe HE 2007 Suppression of long-branchattraction artefacts in the animal phylogeny using a site-heterogeneous model BMC Evol Biol 7S4

Lartillot N Lepage T Blanquart S 2009 PhyloBayes 3 a Bayesian softwarepackage for phylogenetic reconstruction and molecular datingBioinformatics 252286ndash2288

Leite YLR Costa LP Loss AC Rocha RG Batalha-Filho H Bastos ACQuaresma VS Fagundes V Paresque R Passamani M et al 2016Neotropical forest expansion during the last glacial period challengesrefuge hypothesis Proc Natl Acad Sci USA 1131008ndash1013

Leite YLR Patton JL 2002 Evolution of South American spiny rats(Rodentia Echimyidae) the star-phylogeny hypothesis revisitedMol Phylogenet Evol 25455ndash464

Lepage T Bryant D Philippe HE Lartillot N 2007 A general comparisonof relaxed molecular clock models Mol Biol Evol 242669ndash2680

Loss AC Loss Moura RT Leite YLR 2014 Unexpected phylogeneticrelationships of the painted tree rat Callistomys pictus (RodentiaEchimyidae) Natureza 12132ndash135

Lovejoy NR Albert JS Crampton WGR 2006 Miocene marine incursionsand marinefreshwater transitions Evidence from Neotropical fishesJ South Am Earth Sci 215ndash13

Lundberg JG Marshall CR Guerrero J Horton B Malabarba MCSLWesselingh FP 1998 Phylogeny and classification of Neotropicalfishes Edipucrs Porto Alegre

Macphee RDE 2005 ldquoFirstrdquo appearances in the Cenozoic land-mam-mal record of the Greater Antilles significance and comparisonwith South American and Antarctic records J Biogeogr32551ndash564

Martini AMZ Fiaschi P Amorim AM Paix~ao JL 2007 A hot-point withina hot-spot a high diversity site in Brazilrsquos Atlantic Forest BiodiversConserv 163111ndash3128

Matzke NJ 2013 Probabilistic historical biogeography new models forfounder-event speciation imperfect detection and fossils allow im-proved accuracy and model-testing Front Biogeogr 4210

Matzke NJ 2014 Model selection in historical biogeography reveals thatfounder-event speciation is a crucial process in island clades FrontBiogeogr 63syu056ndashsyu970

McKenna MC Bell SK 1997 Classification of mammals above the spe-cies level New York Columbia University Press

Meredith RW Janecka JE Gatesy J Ryder OA Fisher CA 2011 Impacts ofthe Cretaceous terrestrial revolution and KPg extinction on mam-mal diversification Science 334521ndash524

Meyer M Kircher M 2010 Illumina sequencing library preparation forhighly multiplexed target capture and sequencing Cold Spring HarbProtoc 65448

Meyer M Stenzel U Hofreiter M 2008 Parallel tagged sequencing on the454 platform Nat Protoc 3267ndash278

Morrone JJ 2014 Biogeographical regionalisation of the neotropical re-gion Zootaxa 37821ndash110

Nascimento FF Lazar A Menezes AN da Matta Durans A Moreira JCSalazar-Bravo J Andrea PSD Bonvicino CR 2013 The role of histor-ical barriers in the diversification processes in open vegetation for-mations during the miocenepliocene using an ancient rodentlineage as a model Fine PVA editor PLoS One 8e61924

Olivares AI Montalvo CI Verzi DH Vucetich MG 2012 Phylogeneticaffinities of the late miocene echimyid daggerPampamys and the age ofThrichomys (Rodentia Hystricognathi) J Mammal 9376ndash86

Parada A Pardi~nas UFJ Salazar-Bravo J DrsquoElıa G Palma RE 2013 Datingan impressive Neotropical radiation Molecular time estimates forthe Sigmodontinae (Rodentia) provide insights into its historicalbiogeography Mol Phylogenet Evol 66960ndash968

Paradis E Claude J Strimmer K 2004 APE analyses of phylogenetics andevolution in R language Bioinformatics 20289ndash290

Pardi~nas UF Geise L Ventura K Lessa G 2016 A new genus forHabrothrix angustidens and Akodon serrensis (RodentiaCricetidae) again paleontology meets neontology in the legacy ofLund Mastozool Neotrop 2393ndash115

Pascual R Jaureguizar EO 1990 Evolving climates and mammal faunas incenozoic South America J Human Evol 1923ndash60

Patterson BD Costa LP 2012 Bones clones and biomes ChicagoUniversity of Chicago Press

Patterson BD Upham NS 2014 A newly recognized family from theHorn of Africa the Heterocephalidae (Rodentia Ctenohystrica) ZoolJ Linn Soc 172942ndash963

Patterson BD Velazco PM 2008 Phylogeny of the rodent genus isothrix(Hystricognathi Echimyidae) and its diversification in Amazonia andthe Eastern Andes J Mammal Evol 15181ndash201

Patterson BD Solari S Velazco PM 2012 The role of the Andes in thediversification and biogeography of Neotropical mammals Bonesclones and biomes the history and geography of recent neotropicalmammals Chicago University of Press p 351ndash378

Patton JL da Silva MNF Malcolm JR 2000 Mammals of the Rio Juruaand the evolutionary and ecological diversification of Amazonia BullAm Mus Nat Hist Bull 2441ndash306

Patton JL Pardi~nas UFJ DrsquoElıa G 2015 Mammals of South AmericaVolume 2 Chicago University of Chicago Press

Pellegrino KCM Rodrigues MT James Harris D Yonenaga-Yassuda YSites JW Jr 2011 Molecular phylogeny biogeography and insightsinto the origin of parthenogenesis in the Neotropical genusLeposoma (Squamata Gymnophthalmidae) Ancient links betweenthe Atlantic Forest and Amazonia Mol Phylogenet Evol 61446ndash459

Pessoa LM Williams TL Tavares WC Oliveira JA Patton JL 2015Mammals of South America Volume 2 Rodents Chicago ILUniversity of Chicago Press

Philippe HE Lartillot N Brinkmann H 2005 Multigene analyses of bilat-erian animals corroborate the monophyly of EcdysozoaLophotrochozoa and Protostomia Mol Biol Evol 221246ndash1253


631

Dow


icoupcomm


Poe S 2003 Evaluation of the strategy of long-branch subdivision toimprove the accuracy of phylogenetic methods Front Biogeogr52(3) 423ndash428

Ranwez V Delsuc F Ranwez S Belkhir K Tilak M-K Douzery EJP 2007OrthoMaM a database of orthologous genomic markers for placen-tal mammal phylogenetics BMC Evol Biol 7241

Ranwez V Harispe S Delsuc F Douzery EJP 2011 MACSE MultipleAlignment of Coding SEquences accounting for frameshifts andstop codons PLoS One 6e22594

Reuroaseuroanen ME Linna AM Santos J Negri FR 1995 Late Miocene tidaldeposits in the Amazonian foreland basin Science 26938

Reuroaseuroanen ME Linna AM 1996 Miocene deposits in the Amazonianforeland basin Technical comments Science 273122ndash122

Ree RH Moore BR Webb CO Donoghue MJ 2005 A likelihood frame-work for inferring the evolution of geographic range on phylogenetictrees Evolution 592299ndash2311

Ree RH Smith SA 2008 Maximum likelihood inference of geographicrange evolution by dispersal local extinction and cladogenesis FrontBiogeogr 574ndash14

Revell LJ 2012 phytools an R package for phylogenetic comparativebiology (and other things) Methods Ecol Evol 3217ndash223

Roddaz M Viers J Brusset S Baby P Boucayrand C 2006 Controls onweathering and provenance in the Amazonian foreland basinInsights from major and trace element geochemistry of NeogeneAmazonian sediments Chem Geol 22631ndash65

Rodriguez F Oliver JL Marin A Medina JR 1990 The general stochasticmodel of nucleotide substitution J Theor Biol 142485ndash501

Ron SR 2000 Biogeographic area relationships of lowland Neotropicalrainforest based on raw distributions of vertebrate groups Biol J LinnSoc 71379ndash402

Rowe KC Singhal S Macmanes MD Ayroles JF Morelli TL Rubidge EMBi KE Moritz CC 2011 Museum genomics low-cost and high-ac-curacy genetic data from historical specimens Mol Ecol Res111082ndash1092

Salazar-Bravo J Pardi~nas UFJ DrsquoElıa G 2013 A phylogenetic appraisal ofSigmodontinae (Rodentia Cricetidae) with emphasis on phyllotinegenera systematics and biogeography Zool Scr 42250ndash261

Schultz PH Zarate M Hames WE King J Heil C Koeberl C Renne PRBlasi A 2002 Argentine impact record GSA Spec Pap 1786

Sempere T Herail G Oller J Bonhomme MG 1990 Late oligocene-earlymiocene major tectonic crisis and related basins in Bolivia Geology18946ndash949

Swofford DL 2003 PAUP Phylogenetic analysis using parsimony (andother methods) Version 4 Sunderland MA Sinauer Associates

Thorne JL Kishino H Painter IS 1998 Estimating the rate of evolution ofthe rate of molecular evolution Mol Biol Evol 151647ndash1657

Thorne JL Kishino H 2002 Divergence time and evolutionary rate esti-mation with multilocus data Syst Biol 51689ndash702

Tilak M-K Justy F Debiais-Thibaud M Botero-Castro F Delsuc FDouzery EJP 2015 A cost-effective straightforward protocol for shot-gun Illumina libraries designed to assemble complete mitogenomesfrom non-model species Conserv Genet Resour 737ndash40

Tomasco IH Tomasco IH Lessa EP 2011 The evolution of mitochondrialgenomes in subterranean caviomorph rodents Adaptation against abackground of purifying selection Mol Phylogenet Evol 6164ndash70

Turvey ST Oliver JR Narganes Storde YM Rye P 2007 Late Holoceneextinction of Puerto Rican native land mammals Biol Lett3193ndash196

Upham NS 2014 Ecological diversification and biogeography in theNeogene evolution of a major lineage of American and Caribbeanrodents (Caviomorpha Octodontoidea) PhD Thesis p 1ndash272

Upham NS Patterson BD 2012 Molecular phylogenetics and evolutionMol Phylogenet Evol 63417ndash429

Upham NS Patterson BD 2015 Phylogeny and evolution of caviomorphrodents a complete timetree for living genera In Vassallo AIAntenucci D editors Biology of caviomorph rodents diversity andevolution Sociedad Argentina para el Estudio de los Mamıferos(SAREM) Argentina Buenos Aires p 63ndash120

Upham NS Ojala-Barbour R Brito M J Velazco PM Patterson BD 2013Transitions between Andean and Amazonian centers of endemismin the radiation of some arboreal rodents BMC Evol Biol 13191

Ventura K Ximenes GEI Pardini R de Sousa MAN Yonenaga-Yassuda YSilva MJ 2008 Karyotypic analyses and morphological comments onthe endemic and endangered Brazilian painted tree rat Callistomyspictus (Rodentia Echimyidae) Genet Mol Biol 31697ndash703

Verzi DH Olivares AI Morgan CC Alvarez A 2015 Contrasting phylo-genetic and diversity patterns in octodontoid rodents and a newdefinition of the family Abrocomidae J Mammal Evol 2393ndash115

Verzi DH 2008 Phylogeny and adaptive diversity of rodents of the familyCtenomyidae (Caviomorpha) delimiting lineages and genera in thefossil record J Zool 274386ndash394

Verzi DH 2013 Phylogeny evolutionary patterns and timescale of SouthAmerican octodontoid rodents The importance of recognising mor-phological differentiation in the fossil record Acta PalaeontologPolon 4757ndash769

Vilela JF Voloch CM Loss-Oliveira L Schrago CG 2013 Phylogeny andchronology of the major lineages of New World hystricognath ro-dents insights on the biogeography of the EoceneOligocene arrivalof mammals in South America BMC Res Notes 6160

Vilela RV Machado T Ventura K Fagundes V Silva MJ Yonenaga-Yassuda Y 2009 The taxonomic status of the endangered thin-spined porcupine Chaetomys subspinosus (Olfers 1818) based onmolecular and karyologic data BMC Evol Biol 929

Villagomez D Spikings R Mora A Guzman G Ojeda G Cortes E van derLelij R 2011 Vertical tectonics at a continental crust-oceanic plateauplate boundary zone fission track thermochronology of the SierraNevada de Santa Marta Colombia Tectonics 304

Voloch CM Vilela JF Loss-Oliveira L Schrago CG 2013 Phylogeny andchronology of the major lineages of New World hystricognathrodents insights on the biogeography of the EoceneOligocenearrival of mammals in South America BMC Res Notes 61

Voss RS Myers P (1991) Pseudoryzomys simplex (Rodentia Muridae)and the significance of Lundrsquos collections from the caves of LagoaSanta Brazil In Griffiths TA Klingener D editors Contributions tomammalogy in honor of Karl F Koopman New York Bulletin of theAmerican Museum of Natural History p 414ndash432

Vucetich MG Mazzoni MM Pardi~nas UFJ 1993 Los roedores de laFormacion Collon Cura (Mioceno Medio) y la IgnimbritaPilcaniyeu Ca~nadon Del Tordillo Neuquen Ameghiniana30361ndash381

Vucetich MG Vieytes EC Perez ME 2010 The rodents from La Canteraand the early evolution of caviomorphs in South America InMadden RH Carlini AA Vucetich MG kay RF editors The paleon-tology of Gran Barranca evolution and environmental changethrough the Middle Cenozoic of Patagonia UK CambridgeUniversity Press p 193ndash205

Webb SD 1995 Biological implications of the middle Miocene Amazonseaway Science 269361

Weksler M 2009 Phylogenetic relationships of oryzomine rodents(Muroidea Sigmodontinae) separate and combined analyses ofmorphological and molecular data Bull Am Mus Nat Hist 2961ndash149

Wesselingh FP Hoorn MC Guerrero J Reuroaseuroanen ME 2006 The stratig-raphy and regional structure of Miocene deposits in westernAmazonia (Peru Colombia and Brazil) with implications for lateNeogene landscape evolution Scr Geol 133291ndash322

Wesselingh FP Salo JA 2006 A Miocene perspective on the evolution ofthe Amazonian biota Scr Geol 133439ndash458

Wilson DE Lacher TE Jr Mittermeier RA 2016 Handbook of theMammals of the World Vol 6 Lagomorphs and Rodents IBarcelona Lynx Edicions

Wilson DE Reeder DM 2005 Mammal species of the world a taxonomicand geographic reference Vol 2 John Hopkins University Press

Woods CA Borroto Paez R Kilpatrick CW 2001 Insular patterns andradiations of west indian rodents In Woods CA Sergile FE editorsBiogeography of the West Indies patterns and perspectives BocaRaton LA CRC Press p 335ndash353


632

Dow


icoupcomm


Woods CA Contreras L Whidden HP Willner-Chapman G 1992Myocastor coypus Mammal Species 3981ndash8

Woods CA Howland EB 1979 Adaptive radiation of capromyid rodentsanatomy of the masticatory apparatus J Mammal 6095ndash116

Woods CA Kilpatrick CW 2005 Mammal species of the world a taxo-nomic and geographic reference Hystricognathi In Wilson DEReeder DM editors Vol 1 Baltimore John Hopkins University Press

Wyss AR Flynn JJ Norell MA Swisher CCIII Charrier R Novacek MJMcKenna MC 1993 South Americarsquos earliest rodent and recogni-tion of a new interval of mammalian evolution Nature 365434ndash437

Zachos J Pagani M Sloan L Thomas E Billups K 2001 Trends rhythmsand aberrations in global climate 65 Ma to present Science292686ndash693


633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7

Mitogenomic Phylogeny Diversification and Biogeography ofSouth American Spiny Rats

Pierre-Henri Fabre12 Nathan S Upham34 Louise H Emmons2 Fabienne Justy1 Yuri LR Leite5 AnaCarolina Loss5 Ludovic Orlando67 Marie-Ka Tilak1 Bruce D Patterson4 and Emmanuel JP Douzery1

1Institut des Sciences de lrsquoEvolution (ISEM UMR 5554 CNRS-IRD-UM) Universite de Montpellier Montpellier France2National Museum of Natural History Smithsonian Institution Washington DC3Ecology and Evolutionary Biology Yale University New Haven CT4Integrative Research Center Field Museum of Natural History Chicago IL5Departamento de Ciencias Biologicas Universidade Federal do Espırito Santo Vitoria Espırito Santo Brazil6Centre for GeoGenetics Natural History Museum of Denmark University of Copenhagen Copenhagen Denmark7Laboratoire drsquoAnthropobiologie Moleculaire et drsquoImagerie de Synthese Universite de Toulouse University Paul Sabatier ToulouseFrance

Corresponding author E-mail phfmouradegmailcom

Associate editor Emma Teeling

Abstract

Echimyidae is one of the most speciose and ecologically diverse rodent families in the world occupying a wide range ofhabitats in the Neotropics However a resolved phylogeny at the genus-level is still lacking for these 22 genera of SouthAmerican spiny rats including the coypu (Myocastorinae) and 5 genera of West Indian hutias (Capromyidae) relativesHere we used Illumina shotgun sequencing to assemble 38 new complete mitogenomes establishing Echimyidae andCapromyidae as the first major rodent families to be completely sequenced at the genus-level for their mitochondrialDNA Combining mitogenomes and nuclear exons we inferred a robust phylogenetic framework that reveals severalnewly supported nodes as well as the tempo of the higher level diversification of these rodents Incorporating the fullgeneric diversity of extant echimyids leads us to propose a new higher level classification of two subfamiliesEuryzygomatomyinae and Echimyinae Of note the enigmatic Carterodon displays fast-evolving mitochondrial andnuclear sequences with a long branch that destabilizes the deepest divergences of the echimyid tree thereby challengingthe sister-group relationship between Capromyidae and Euryzygomatomyinae Biogeographical analyses involving higherlevel taxa show that several vicariant and dispersal events impacted the evolutionary history of echimyids The diver-sification history of Echimyidae seems to have been influenced by two major historical factors namely (1) recurrentconnections between Atlantic and Amazonian Forests and (2) the Northern uplift of the Andes

Key words biogeography diversification Echimyidae mitogenomics Neotropics nuclear DNA fast-evolving geneCarterodon

Introduction

Thanks to a unique geological history and ecology theNeotropical biota together composes the richest realm ofbiodiversity in the world The formation of complex riverinesystems and the uplift of the Andean Cordillera has togetherwith climatic changes driven several spectacular biologicalradiations (Hoorn et al 2010 Patterson and Costa 2012)Based on geology and zoogeography (Morrone 2014) thetropical Americas can be subdivided into six major biomes(fig 1) Amazon Basin Central Andes and Guyana Shield(AM) the trans-Andean region plus Choco (CH) theNorthern Andes and northern coastal dry forests inVenezuela and Colombia (NA) Caatinga Chaco Cerradoand the Pantanal (CC) Atlantic Forest (AF) and the insularWest Indies (WI) These biomes are characterized by severallarge forested regions (Amazon Basin Atlantic Forest ChocoOrinoco watersheds and West Indies) open habitats (CC

CerradothornCaatinga and Chaco NA Pantanalthorn Llanos)and the Andean mountain range (CentralthornNorthernAndes) The major barriers currently separating the forestedbiomes were formed following climate aridification andormountain uplift during the Miocene (Hoorn et al 2010)and have substantially contributed to biotic divergence andendemism in the Neotropics The Amazon Basin forms thelargest Neotropical biome and is surrounded to the west bythe Choco wet forest (across the Northern Andes) both theOrinoco watershed and the northern coastal dry forests inVenezuela and Colombia and the coastal Atlantic Forest ofBrazil and Paraguay Disjunct distributions of terrestrial ver-tebrate species among the Amazon Basin and the AtlanticForests reflect a common geoclimatic divergence structuringin South American biotas (da Silva and Patton 1993 Pattonet al 2000 Batalha-Filho et al 2013) Biotas of those tworainforests are among the most diverse globally and are sep-arated by a diagonal region of xeric habitats comprising the

Article

The Author 2016 Published by Oxford University Press on behalf of the Society for Molecular Biology and EvolutionAll rights reserved For permissions please e-mail journalspermissionsoupcom

Mol Biol Evol 34(3)613ndash633 doi101093molbevmsw261 Advance Access publication December 26 2016 613

Dow


icoupcomm







614

Dow


icoupcomm











615

Dow


icoupcomm











616

Dow


icoupcomm











617

Dow


icoupcomm











618

Dow


icoupcomm







619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7






614

Dow


icoupcomm











615

Dow


icoupcomm











616

Dow


icoupcomm











617

Dow


icoupcomm











618

Dow


icoupcomm







619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7










615

Dow


icoupcomm











616

Dow


icoupcomm











617

Dow


icoupcomm











618

Dow


icoupcomm







619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7










616

Dow


icoupcomm











617

Dow


icoupcomm











618

Dow


icoupcomm







619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7










617

Dow


icoupcomm











618

Dow


icoupcomm







619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7










618

Dow


icoupcomm







619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7






619

Dow


icoupcomm









620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7








620

Dow


icoupcomm













621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7












621

Dow


icoupcomm








622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7







622

Dow


icoupcomm










623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7









623

Dow


icoupcomm








624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7







624

Dow


icoupcomm












625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7











625

Dow


icoupcomm







626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7






626

Dow


icoupcomm









627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7








627

Dow


icoupcomm








Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7







Classification




9 8 (88)







46 35 (76)



628

Dow


icoupcomm



Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7


Taxonomy














629

Dow


icoupcomm















































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7














































630

Dow


icoupcomm
















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7















































631

Dow


icoupcomm














































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7













































632

Dow


icoupcomm








633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7







633

Dow


icoupcomm


msw261-TF1

msw261-TF2

msw261-TF3

msw261-TF4

msw261-TF5

msw261-TF6

msw261-TF7

mitogenomic phylogeny, diversification, and biogeography

Documents