Phylogenetic analyses of Postcranial skeletal Morphology in Didelphid Marsupials

Abstract

In this study I provide a phylogenetic hypothesis for didelphid Marsupials including a suite of 114 postcranial characters. The postcranial evidence was cladistically analyzed separately and concatenated with a nonmolecular data set previously published (71 cranio-dental-external characters). A combined analysis was done including published IRBP (interphotoreceptor retinoid binding protein, 1158 bp), DMP-1 (dentin matrix protein 1, 1176 bp), and RAG-1 (recombinase-activating gene, 2790 bp) sequences to the nonmolecular data set. In order to compare and evaluate the influence of the inclusion of postcranial morphology to previous hypotheses, the taxon sampling of didelphine ingroup was almost similar to the one used in recent series of papers on didelphid phylogeny. The postcranial information includes 48 characters from the axial skeleton, 37 from the anterior limb, and 29 from the posterior limb. I present anatomical descriptions for each postcranial character, adding details of different conditions observed among didelphine ingroup, as well as some functional implications. Different hypotheses that are discussed as polymorphic characters are alternatively treated as composite entries (CO) and transformation series (TS) in morphological and combined analyses. Different codings of polymorphic postcranial characters produce topologies that in general are not contradictory. The principal difference is the loss of resolution of trees in TS analysis, compared to CO analysis in postcranial evidence, whereas the support values were in general low in both codings. The topology obtained from postcranial evidence supported some already recovered relationships, such as the monophyly of the large opossums (Didelphis, Philander, Chironectes, Lutreolina, and Metachirus), and several polytypic groups such as Didelphis, Monodelphis, Marmosops, Thylamys, Micoureus, and Philander. Additionally, the intermediate position of Hyladelphys between calorumyines and didelphines is kept in CO analysis. The inclusion of the postcranial data set to previous nonmolecular evidence causes little incongruence, although some modifications in the topology and support values were detected. The effect of different codings of polymorphic characters was similar respect to the postcranialonly data set. In this case, the topology obtained with CO analysis was also notably better resolved than TS analysis. Similarly to the postcranial-only analysis, the topologies obtained in the total morphological evidence applying the two kinds of codings are highly congruent, but the TS treatment seemed not to contribute to retention of more phylogenetic information, since the CO analysis was better resolved. The relationships obtained adding the postcranial evidence to published combined data set (i.e., morphology, IRBP, DMP-1, and RAG-1 sequences) were mostly better resolved and supported in the CO coding than the morphological analyses, although the TS coding causes loss of resolution in the strict consensus. In this sense, some strong differences on deep branch topology can be detected depending on the treatment applied to polymorphic entries and partitioned analyses (e.g., phylogenetic condition of the mouse opossums, nodes C and B). Including the postcranial evidence in the total data set, I also recovered the intermediate position of Hyladelphys, but never the distantly related clades recently recovered by the inclusion of RAG-1 sequences (clades B + I in Gruber, K.F., R.S. Voss, and S.A. Jansa. 2007. Base-compositional heterogeneity in the RAG1 locus among didelphid marsupials: implications for phylogenetic inference and the evolution of GC content. Systematic Biology 56: 1–14). However, the position of Metachirus nudicaudatus and Tlacuatzin canescens is highly affected. Excluding the last gene, the partitioned combined analysis considering postcranial evidence was highly congruent with previous IRBP, morphology, and combined topologies, especially in the TS analysis, where all already defined nodes were recovered. The inclusion of postcranial evidence to the previous combined data set actually improves the support values when RAG-1 is eliminated. However, even when the topologies from both kinds of codings of polymorphic data were considerably congruent, the mouse opossums were clustered in CO analysis (although including Metachirus if RAG-1 sequences are included). The application of different criteria for the treatment of polymorphic data affects the relationships and phyletic condition of the mouse opossums. Finally, I discuss the postcranial morphology as evidence of didelphid phylogeny, as well as the new postcranial synapomorphies found in the monophyletic groups recovered in the didelphid crown group.