Análise filogenética de Gymnophiona Müller, 1832 (Amphibia)

Abstract

Gymnophiona Müller, 1832—with 206 currently recognized species—is the least studied order within Amphibia. The phylogenetic relationships among its members have been historically unstable, with frequent taxonomic changes at the family level due to the recurrent presence of paraphyletic taxa. However, an increase of studies based on morphological and molecular data has built a scaffold of information about the evolutionary relationships among caecilians; even though, a phylogenetic classification of the order has remained a challenge. In 2011 a new taxonomy of Gymnophiona was proposed, where genera were arranged in nine families supposedly monophyletic. This taxonomy was not based on a phylogenetic analysis but on a consensus of the understanding of the evolutionary relationships of the group as inferred by previous studies. However, important conflicts exits among the results and type of analyses performed on these studies. Furthermore, only one study has included the known fossil taxon of Gymnophiona, which despite being fragmented is a valid source of evidence. Due to several and continuous efforts to generate data on caecilians, there is a wealth of hereditary characters available for phylogenetic studies. Thus, it has become crucial to perform a combined analysis of all these data (complete mitochondrial genomes, nuclear DNA sequences, phenotypic characters of extant and fossil taxa) to obtain a phylogenetic hypothesis that maximizes explanatory power. The objective of this study is to infer the evolutionary relationships of Gymnophiona based on a total evidence analysis using parsimony and dynamic homology to evaluate the current taxonomy of the group. Additionally, we also evaluate the effect of different types of alignment (tree versus similarity approaches), the influence of phenotypic characters on a dataset dominated by molecular characters, and the effect of coding indels as missing data. We compiled previously published phenotypic characters, DNA sequences available from GenBank of 47 nuclear and mitochondrial genes of all available taxa, and produced 42 new sequences. When comparing the results obtained from the tree-alignment analysis in POY with those of the similarity alignment in TNT, both recover Rhinatrematidae as sister of Ichthyophiidae + Teresomata. Within Rhinatrematidae, Epicrionops is paraphyletic in all analyses and the relationships within Ichthyophis are unsolved (a polytomy in the strict consensus). POY does not recovered Scolecomorphidae as the sister taxon of all other Teresomata but Typhlonectidae + Caeciliidae. Caecilia is paraphyletic with respect to Oscaecilia and Typhlonectes in relationship to Potamotyphlus. Scolecomorphidae is sister of Herpelidae + Chikilidae. The results of TNT recover a paraphyletic Herpelidae, with Herpele squalostoma sister of Chikilidae. Also, Siphonopidae is non-monophyletic. Indotyphlidae is non-monophyletic in both analyses and Idiocranium is consistently recovered as sister taxon of Dermophiidae. Dermophis is recovered as paraphyletic in the POY analysis, while both Dermophis and Gymnopis are paraphyletic in the TNT analysis. The strict consensus of the molecular dataset is highly congruous with that of the total evidence dataset; however, the former is better resolved (less polytomies), mainly within Indotyphlidae. The analysis of the phenotypic data alone resulted in a complete polytomy, illustrating the need of more research in this avenue. Coding indels as missing data did not cause important topological changes. The main conclusions derived from this study are: (i) the type of alignment of DNA sequences have an evident impact on the phylogenetic hypotheses of Gymnophiona; (ii) the apparent resolution on the evolutionary relationships of the extant supraspecific taxa of Gymnophiona and their monophyly presented in other studies are dependent on the exclusion of relevant evidence—taxa and characters—or the partial presentation of the optimal hypotheses; (iii) only a total evidence analysis allowed us to discover some of the potential cases of paraphyly or misidentification of vouchers; (iv) the phenotypic data currently used in the study of the evolutionary relationships of Gymnophiona contain important levels of non-congruent information and are not sufficient to place the fossil Eocaecilia micropodia within caecilians. This study reveals the need of detailed revision of the taxonomy and phylogeny of Gymnophiona.