Mammals from ‘down under’: a multi-gene species-level phylogeny of marsupial mammals (Mammalia, Metatheria)

Paucituberculata

Within Paucituberculata, the genus Caenolestes was sister to a clade containing Lestoros and Rhyncholestes was strongly supported in all concatenated and mtDNA analyses (Figs. S1 and S4).

Didelphimorphia

Within Didelphimorphia, the monophyly of subfamily Caluromyinae as recognized by Voss & Jansa (2009) containing Caluromys and Caluromysiops was supported (Figs. S1–S3). The only extant species of the subfamily Hyladelphinae (Voss & Jansa, 2009), the Kalinowski’s mouse opossum (Hyladelphys kalinowskii), was sister to Didelphinae, and the only member of the subfamily Glironiinae, the bushy-tailed opossum; Glironia venusta was placed in most analyses basal to the clade containing the Hyladelphinae and the Didelphinae (Fig. 2 and Figs. S1–S3). The monophyly of the subfamily Didelphinae (Marmosa, Monodelphis, Tlacuatzin, Metachirus, Chironectes, Didelphis, Lutreolina, Philander and Chacodelphys, Cryptonanus, Gracilinanus, Lestodelphys, Marmosops, and Thylamys) was supported by the all analyses except the mtDNA partition (Fig. S4). The monophyly of the genera Thylamys, Cryptonatus, Marmosops, Monodelphis, Gracilinanus, Philander and Marmosa (plus subgenus Micoureus) were supported by most analyses (Fig. 2 and Figs. S1–S6). Other interesting relationships include the sister relationship of the genera Cryptonanus and Gracilinanus, the placement of the grayish mouse opossum (Tlacuatzin canescens) sister to Marmosa, and the close relationship of the Patagonian opossum (Lestodelphys halli) with members of the genus Thylamys, and the basal position of the Virginia opossum (Didelphis virginiana) to Philander and the other taxa of Didelphis (Fig. 2 and Figs. S1–S6).

Peramelemorphia

Within Peramelemorphia, the monophyly of the family Peramelidae was corroborated in three analyses (Fig. 1). The extinct member of the family Chaeropodidae the pig-footed bandicoot (Chaeropus ecaudatus) was placed as sister to the remaining peramelemorphians in the full multigene and mtDNA analyses with strong support (Figs. S1 and S4). In contrasts, it was placed in the nuDNA partition within the Peramelidae either sister to Echymiperinae (Fig. 3) or within the Microperoryctes (Figs. S2–S3) or within Diprotodontia (Fig. S6) but the support for these placements was low. In six of the analyses the greater bilby, Macrotis lagotis (Thylacomyidae) was placed as sister to the remaining peramelemorphians (Fig. 3 and Figs. S1, S4, S6). The monophyly of the genera Echymipera, Peroryctes, Isoodon and Perameles was supported by all analyses. The concatenated analyses differed from the mtDNA partition in the relations among genera. The most supported relationship was ((Echymipera + Microperoryctes) Peroryctes)+(Isoodon + Perameles) (Fig. 3 and Figs. S1 and S6).

Dasyuromorphia

Within Dasyuromorphia, the monophyly of the families Dasyuridae, Myrmecobiidae, and Thylacinidae was supported by most analyses with the interrelationships (Thylacinidae (Myermecobiidae, Dasyuridae)) (Fig. 1). In two analyses, the extinct Tasmanian wolf (Thylacinus cynocephalus) was placed within the subfamily Dasyurinae, but this relationship was poorly supported (Figs. S2 and S3). The monophyly of the two Dasyuridae subfamilies Dasyurinae and Sminthopsinae was also supported by most analyses, as well as the two Dasyurinae tribes Dasyurini and Phascogalini and the two Sminthopsinae tribes Planigalini and Sminthopsini (Fig. 3 and Figs. S1–S4). In most of the analyses, the phylogenetic position of Ningaui and the Kultarr (Antechinomys laniger) rendered the genus Sminthopsis paraphyletic (Fig. 3 and Figs. S1–S4). Also, highly consistent across analyses is the phylogenetic position of the Mulgara (Dasycercus cristicauda) as sister to the Ningbing fals antechinus, Pseudantechinus ningbing also rendering the later paraphyletic (Fig. 3 and Figs. S1–S4). The monophyly of the genera Planigale, Murexia, Antechinus, Phascogale, Myoictis, and Dasyurus was supported by most analyses (Fig. 3 and Figs. S1–S4).

Diprotodontia

With the exception of Phalangeriformes, the monophyly of other suborders, Vombatiformes (Vombatidae + Phascolarctidae) and Macropodiformes (Hypsiprymnodontidae + Macropodidae + Potoroidea) were supported by all analyses. The monophyly of superfamily Macropodidea was also supported (Fig. 1). The position of Vombatiformes as sister to the remaining Diprododontia was supported by all analyses except the mtDNA partition (Fig. 4 and Figs. S1–S4). Within the superfamily Petauroidea (Acrobatidae, Tarsipedidae, Pseudocheiriidae, and Petauridae), all partitions except mtDNA supported the position of the family Acrobatidae as the basal petauroids (Fig. 4 and Figs. S1–S3). Tarsipedidae was sister to a clade containing Petauridae and Pseudocheiridae. Within Petauridae, the placement of the Gymnobelideus sister to the genus Dactylopsila to the exclusion of Petaurus was strongly supported by both concatenated and the mtDNA partitions (Fig. 4 and Figs. S1–S4). The monophyly of the Pseudocheiridae was strongly supported, as well as that of the subfamilies Pseudocheirinae (Pseudochirulus, Pseudocheirus), Hemibelidinae (Hemibelideus, Petauroides) and Pseudochiropsinae (Petropseudes, Pseudochirops). The monophyly of genus Pseudochirops was not supported is most analyses (Fig. 4 and Figs. S1–S4). The superfamily Petauroidea is the sister group of Macropodiformes, and Phalangeroidea is the sister group to the clade containing Phalangeridae and Burramyidae (Fig. 1). Within the Phalangeridae, the Sulawesi cuscus (Strigocuscus celebensis) was the sister species of the Sulawesi bear (Ailurops ursinus) in all partitions, rendering the genus Strigocuscus paraphyletic, and the scaly tailed opossum Wyulda squamicaudata was sister to Trichosurus (Fig. 4 and Figs. S1–S4). The three kangoroo families Potoroidea, Macropodidae, and Hypsiprymnodontidae were monophyletic and their interrelationships Hypsiprymnodontidae (Potoroidea (Macropodidae) were also supported by all partitions except mtDNA (Fig. 4 and Figs. S1–S4). Within Potoroidea, the Rofous bettong, Aepyprymnus rufescens was sister to Bettongia to the exclusion of Potorous (Fig. 4 and Figs. S1–S3). Within Macropodidae, the monophyly of Macropus and Petrogale, was not supported by most analyses (Fig. 4 and Figs. S1–S3). In the full multigene analysis, Bennett’s tree kangaroo, Dendrolagus bennettianus, was the basal species in macropodids (Fig. S1), but for the pruned multigene analysis and nuclear analyses it was instead sister to all Phalanger species (Fig. 4 and Figs. S2 and S3).

Paucituberculata and Didelphimorphia

As found by recent molecular and morphological studies, the two Ameridelphian orders Paucituberculata and Didelphimorphia form a grade rather than a clade (Horovitz & Sánchez-Villagra, 2003; Nilsson et al., 2010). However, we cannot reject the monophyly of Ameridelphia. Our results corroborate Voss & Jansa’s (2009) combined morphological and molecular study where Glirona is placed in a monotypic subfamiliy Glironiinae, whereas Caluromys and Caluromysiops remained in the subfamily Caluromyninae. In addition, Hyladelphys was not grouped with other didelphids, thus we support its placement within a subfamily of its own Hyaldephinae (Voss & Jansa, 2009). However, not all analyses supported Didelphinae tribes (Marmosini, Metachirini, Didelphini, and Thylamyini). The species of the genus Cryptonanus (Voss, Lunde & Jansa, 2005) were formerly included in the genus Gracilinanus by Gardner & Creighton (1989). Previous studies have found it difficult to establish the phylogenectic position of Cryptonanus (Jansa, Forsman & Voss, 2006; Flores, 2009); however, with the exception of the mtDNA analyses, we found strong support for a sister relationship between Cryptonanus and Gracilinanus. A close phylogenetic position of the greyish mouse opossum, Tlacuatzin canescens (formely Marmosa canescens) with species of the genus Marmosa (and subgenus Micoureus) is also supported, as has been found in previous studies combining molecules and morphology (Flores, 2009; Voss & Jansa, 2009). The basal placement of the Peruvian gracile mouse opossum, Hyladelphys kalinowskii in the Didelphinae, is also supported by morphological and molecular analyses (Flores, 2009), as well as the close phylogenetic relationship between Patagonian opossum Lestodelphys halli and the genus Thylamys (Cardillo et al., 2004; Jansa, Forsman & Voss, 2006; Flores, 2009).

Dasyuromorphia

Our results for Dasyuromorphia are largely congruent with previous studies in supporting the monophyly of dasyuromorphian families and subfamilies. The phylogenetic position of the Numbat (Myrmecobius fasciatus) and the Tasmanian wolf (Thylacinus cynocephalus) are traditionally controversial. In this study, most analyses strongly supported the basal placement of the Tasmanian wolf and the Numbat to all dasyuromorphians. This phylogenetic arrangement has been previously supported by morphological (Wroe et al., 2000) and combined analyses (Beck, 2008) including a supertree analysis (Cardillo et al., 2004). The position of the Ningaui, and the Kultarr (Antechinomys laniger) within Sminthopsis agrees with previous studies (Archer, 1975; Blacket et al., 1999; Cardillo et al., 2004) and suggests that taxonomic changes are necessary to accommodate this phylogenetic perspective, In the supertree analysis by Cardillo et al. (2004), the Kultarr was referred as Sminthopsis laniger and was placed as basal to a clade containing Ningaui and Sminthopsis species. Also highly consistent across analyses in this study is the phylogenetic position of the Mulgara, Dasycercus cristicauda as sister to the Ningbing false antechinus, Pseudantechinus ningbing, contrasting the results by Cardillo et al. (2004).

Peramelemorphia

The results from the complete multigene and mtDNA data sets are congruent with recent studies placing the extinct pig-footed bandicoot (Chaeropus ecaudatus) sister to other bandicoots (Cardillo et al., 2004; Westerman et al., 2012). Our studies also support a proposed early divergence of the greater bilby, Macrotis lagotis, from other bandicoots (Cardillo et al., 2004; Meredith, Westerman & Springer, 2008a; Meredith, Westerman & Springer, 2008b; Westerman, Meredith & Springer, 2010). The greater bilby is known to be drastically different in their genetic make-up, in that it possesses sixteen autosomes and a multiple sex-chromosome system (Westerman, Meredith & Springer, 2010), contrasting with the typical bandicoot chromosome set of 2n = 14 (Westerman et al., 2012). In addition, our results support the close phylogenetic relationships between the two New Guinean subfamilies Echymiperinae and Peroryctinae (Meredith, Westerman & Springer, 2008a; Meredith, Westerman & Springer, 2008b; Westerman et al., 2012).

Diprotodontia

Like previous studies, our analysis strongly supported the monophyly of Diprotodontia. With the exception of Phalangeriformes the monophyly of the suborders Vombatiformes and Macropodiformes was supported contrasting with one molecular study (Meredith, Westerman & Springer, 2009). Amrine-Madsen et al. (2003) found ambivalent support for the monophyly of Phalangeriformes, and other studies have been inconclusive (Springer & Kirsch, 1991; Kirsch, Lapointe & Springer, 1997; Burk et al., 1999). The basal position of Vombatiformes (Vombatidae + Phascolarctidae) is in agreement with previous molecular studies (Meredith, Westerman & Springer, 2009; Nilsson et al., 2010).

The monophyly of the superfamily Phalangeroidea and families Phalangeridae and Burramyidae was supported and in accordance with previous studies (Osborne, Christidis & Norman, 2002; Beck, 2008; Phillips & Pratt, 2008; Meredith, Westerman & Springer, 2009). Our results are also congruent with previous findings supporting the clade Petauroidea where the families Petauridae and Pseudocheiridae grouped together to the exclusion of Tarsipedidae and Acrobatidae (Osborne, Christidis & Norman, 2002; Kavanagh et al., 2004; Phillips & Pratt, 2008; Meredith, Westerman & Springer, 2009; Meredith et al., 2010). Within the family Petuaridae, our results contrast recent molecular studies and previous morphological studies, in that the genus Gymnobelideus grouped with to the genus Dactylopsila to the exclusion of Petaurus (Smith, 1984; Aplin & Archer, 1987; Springer, Westerman & Kirsch, 1994; Meredith et al., 2010). These results are more in accordance with the molecular analysis of Edwards & Westerman (1995). Like a previous study (Meredith et al., 2010), we find strong support for the monophyly of the family Pseudocheiridae and three subfamilies. Our results agree with Meredith et al. (2010) in that the genus Pseudochirops is paraphyletic due to the placement of the Australian rock-haunting ringtail possum Petropseudes dahli with New Guinean species of Pseudochirops. Our results are also in agreement with Meredith et al. (2010), in the genera placed in the clades corresponding to the Pseudocheirinae and Pseudochiropsinae in contrast to the assignments by Groves (2005).

Petauroidea grouped with Macropodiformes to the exclusion of Phalangeroidea, contrasting previous studies results (Phillips & Pratt, 2008; Meredith, Westerman & Springer, 2009). As shown by previous studies, the three kangaroo families are monophyletic (Osborne, Christidis & Norman, 2002; Meredith, Westerman & Springer, 2009; Phillips et al., 2013). One interesting result was the consistent placement of the Sulawesi cuscus, Strigocuscus celebensis, sister species to the Sulawesi bear, Ailurops ursinus, rendering the genus Strigocuscus paraphyletic. This result agrees with previous molecular studies (Meredith, Westerman & Springer, 2009). Within Macropodiformes, the family Hypsiprymnodontidae was sister to a clade consisting of Macropodidae and Potoroidae. This relationship agrees with a previous study (Burk, Westerman & Springer, 1998). Within Potoroidae the relationship between Bettongia and Aepyprymnus to the exclusion of Potorous was strongly supported similarly to previous support (Burk, Westerman & Springer, 1998). The monophyly of the genus Petrogale was strongly supported by this analysis but phylogenetic relationships among species contrast those proposed by Campeau-Péloquin et al. (2001).