More is needed—Thousands of loci are required to elucidate the relationships of the ‘flowers of the sea’ (Sabellida, Annelida)

Highlights

• Fabriciidae is the sister taxon to a Sabellidae + Serpulidae clade.

• Sabellidae now consists of the subfamilies; Myxicolinae and Sabellinae.

• Myxicolinae contains two tribes, named herein as Amphiglenini and Myxicolini.

• Analysing more genes results in a more stable sabellid phylogeny with higher support.

• Higher divergence genes better resolve the sabellid relationships.

Abstract

Sabellida is a well-known clade containing tube-dwelling annelid worms with a radiolar crown. Iterative phylogenetic analyses over three decades have resulted in three main clades being recognized; Fabriciidae, Serpulidae and Sabellidae, with Fabriciidae proposed as the sister group to Serpulidae. However, relationships within Sabellidae have remained poorly understood, with a proliferation of genera. In order to obtain a robust phylogeny with optimal support, we conducted a large-scale phylogenomic analysis with 19 new sabellid transcriptomes for a total of 21 species. In contrast to earlier findings based on limited DNA data, our results support the position of Fabriciidae as sister taxon to a Sabellidae + Serpulidae clade. Our large sampling within Sabellidae also allows us to establish a stable phylogeny within this clade. We restrict Sabellinae to a subclade of Sabellidae and broaden the previously monotypic Myxicolinae to include Amphicorina and Chone. We tested the robustness of species tree reconstruction by subsampling increasing numbers of genes to uncover hidden support of alternative topologies. Our results show that inclusion of more genes leads to a more stable topology with higher support, and also that including higher divergence genes leads to stronger resolution.

Introduction

Sabellida is a diverse group of tube-dwelling annelids with radiolar crowns. These generally sessile worms use their crown for respiration and to gather suspended food particles from the water column, as the remaining body stays within the tube. The often colorful appearance of these crowns and their rapid withdrawal into the safety of the tubes, have for long attracted scientists and SCUBA divers alike, resulting in their common names fan worms, feather-duster worms and the affectionate general epithet “flowers of the sea” (Jones, 1973) (Fig. 1).

Three family-ranked taxa are grouped within Sabellida: Serpulidae Rafinesque, 1815; Sabellidae Latreille, 1825 and Fabriciidae Rioja, 1923 (see Kupriyanova and Rouse, 2008). Traditionally the clade Sabellida included additional taxa. Sabellida sensu Fauchald, 1977, Knight-Jones, 1981 also contained Sabellariidae, which was based on their shared chaetal inversion. However, Kieselbach and Hausen (2008) showed significant differences in the specific chaetal arrangement of Sabellidae and Sabellariidae, suggesting an independent evolution of this pattern. Phylogenomic data suggests that Sabellariidae forms a clade with Spionidae (Andrade et al., 2015, Helm et al., 2018, Weigert and Bleidorn, 2016). Rouse and Fauchald (1997) also placed Siboglinidae (formerly outside Annelida, as Pogonophora and Vestimentifera) and Oweniidae inside Sabellida. These taxa no longer form a clade with Sabellidae, Fabriciidae and Serpulidae according to molecular data (Andrade et al., 2015, Helm et al., 2018, Weigert and Bleidorn, 2016). Following this evidence, in this study we restrict the name Sabellida to the clade composed of Fabriciidae, Sabellidae and Serpulidae.

The close affinity of serpulids, the calcareous-tube worms, and sabellids (previously including fabriciids) has long been accepted and supported by a variety of morphological features such as the unique kind of chaetal inversion (Kieselbach and Hausen, 2008), an inverted faecal groove, and a radiolar crown (Dales, 1962, Fauchald, 1977, Fitzhugh, 1989). The phylogenetic relationships within Sabellida have been less clear from morphological data matrices and PCR-based sequence data (Brown et al., 1999, Capa et al., 2011a, Colgan et al., 2006, Dales, 1962, Fauchald, 1977, Fitzhugh, 1989, Kupriyanova and Rouse, 2008, Rouse and Fauchald, 1997, Rousset et al., 2007, Rousset et al., 2004, Smith, 1991, Struck et al., 2007, Zrzavý et al., 2009). Though the monophyly of Serpulidae is well supported (Kupriyanova et al., 2006), the former delineation of Sabellidae (including fabriciids) was shown to be paraphyletic based on sequences from three nuclear genes (Kupriyanova and Rouse, 2008). Therefore, Fabriciinae were removed from Sabellidae, and elevated to Fabriciidae and Sabellidae membership was reduced to the previous subfamily Sabellinae (Kupriyanova and Rouse, 2008). Further, the relationships within Sabellidae have not been resolved with confidence (Capa et al., 2011a). Sabellida has been the focus of interest with regards to life history evolution since they show a wide range of reproductive modes and body sizes (Rouse and Fitzhugh, 1994, Kupriyanova et al., 2001, Rouse et al., 2006). However, without a reliable phylogeny for the group, robust inferences about transformations in life history characters cannot be made.

Transcriptomic datasets of hundreds or thousands of genetic loci have recently transformed the understanding of the annelid tree of life (Andrade et al., 2015, Helm et al., 2018, Struck et al., 2015, Weigert et al., 2014, Weigert and Bleidorn, 2016) and have suggested that some relationships cannot be confidently resolved with a small number of loci (Stiller et al., 2020). A phylogenomic analysis is still missing for the three lineages of Sabellida and within Sabellidae.

The main aim of this study is to resolve the phylogeny of Sabellida, first to elucidate the relationships of the three main lineages and with specific focus on the diversity of Sabellidae. Nineteen new sabellid transcriptomes were sequenced for this study resulting in a sequence matrix containing up to 3015 orthologous genes and around one million amino acid positions. We find a topology among the Sabellida lineages that differs from previous analyses based on few genes. We therefore conducted a thorough methodological analysis using up to 13,393 gene trees to test what effect adding more and more gene trees had on the robustness of the new sabellid phylogeny.