Phylogenetic relationships of the family Axinellidae (Porifera: Demospongiae) using morphological and molecular data

Abstract

Alvarez, B., Crisp, M.D., Driver, F., Hooper, J.N.A. & Van Soest, R.W.M. (2000). Phylogenetic relationships of the family Axinellidae (Porifera: Demospongiae) using morphological and molecular data. —Zoologica Scripta, 29, 169–198.Twenty‐seven species of marine sponges belonging to Axinellidae and related groups (Halichondriidae, Dictyonellidae, Agelasida) were selected to test the monophyly of Axinellidae and investigate their phylogenetic relationships using parsimony and maximum likelihood methods. Partial 28S rDNA sequences, including the D3 domain, and traditional morphological characters (mainly skeletal ones) were used independently to construct phylogenetic trees. Sequences were aligned using the appropriate model of secondary structure of the RNA and compared to that produced by the multiple sequence alignment program, ClustalW. The alignment using secondary structure constraints produced a better estimate of the phylogeny and was demonstrated to be an effective and objective method.Results of the cladistic analyses of the molecular and morphological data sets were not fully congruent; the morphological data suggest that Axinellidae is monophyletic, however, the molecular data suggest that it is nonmonophyletic. The single most‐parsimonious tree derived from the molecular data showed that species of Axinella (except A. polypoides) are united in a clade that is more closely related to members of Agelasida than to other species of Axinellidae; the remaining members of Axinellidae form a monophyletic group that is closely related to the families Dictyonellidae and Halichondriidae. The consensus tree of 20 most‐parsimonious trees from the morphological analysis, on the other hand, showed that all the sampled species of Axinellidae belong to a monophyletic group which is closely related to the species of Dictyonellidae and Halichondriidae. Only two branches were identical in both cladograms, the one uniting the species of Ptilocaulis and Reniochalina and the one with the species of Dictyonellidae.The robustness of the molecular and morphological trees (or parts of the trees), was tested using bootstrap, jack‐knife, PTP and T‐PTP tests. The results of the PTP test were significant indicating significant cladistic structure in both data sets. The bootstrap and jack‐knife values indicate that the molecular tree is in general better supported than the morphological one. The lack of morphological characters and the homoplastic nature of some may explain the weak support of the morphological tree. A T‐PTP test of nonmonophyly showed that the nonmonophyly of Axinellidae, as indicated by the results of the molecular analysis, is not significant; however, a T‐PTP test of monophyly of Axinellidae, as indicated by the morphological tree, produced significant results. This indicates that the monophyly of Axinellidae based on morphological data cannot be rejected; the family however, cannot be defined in terms of a unique diagnostic character common to all members of the ingroup.Tests of heterogeneity (reciprocal T‐PTP and partition homogeneity test) indicated that the data partitions are heterogeneous, which could be due to sampling errors (in either data set) or differences in the underlying phylogenies; therefore data were not combined in a single analysis. Further, both data sets are unequally sized (95 informative molecular characters vs. 16 informative morphological characters), which means that the molecular signal could swamp the morphological signal if the data is combined.Nonmonophyly of Axinellidae is supported by chemical and genetic evidence available in the literature and DNA sequences data of axinellid species from New Zealand. However, this needs to be confirmed using independent evidence from different genes (or gene regions), biochemistry, histology or cell ultrastructure. Therefore, no changes to the taxonomic position of the family in the higher classification are proposed at this stage.