Dataset: Cannon_2013_T2: NCBI accessions
Deployment: Halanych_lab_2011-16

from Cannon et al (2013) Biol Bull:

Collection:
Eleven specimens sequenced in this study were collected in September 2011 during Senckenberg’s German Center for Marine Biodiversity Research (DZMB) IceAGE expedition led by Dr. Saskia Brix aboard the R/V Meteor, which circled Iceland, crossing the Mid-Atlantic and Greenland-Scotland ridges. Cephalodiscus specimens were collected during two research expeditions to the Antarctic Peninsula aboard the R/V Lawrence M. Gould in 2001 and 2004. Antarctic enteropneusts were collected in January-February 2013 in the Amundsen and Ross Seas by the R/V Nathaniel B. Palmer. Enteropneusts from Norway were collected on the R/V Håkon Mosby or R/V Aurelia with the aid of Dr. Christiane Todt and the late Dr. Christoffer Schander. Balanoglossus sp. specimens from Mississippi were collected with the assistance of Dr. Richard Heard. Dr. Jon Norenburg and Dr. Darryl Felder provided material from Rhabdopleura sp. collected in the Gulf of Mexico on the R/V Pelican. Specimens were collected at depths ranging from the intertidal to over 2500 m, using diverse sampling methods (Table 1). Enteropneusts from Iceland were collected by decanting sediment through a 1-1.5-mm sieve, and were retained on either a 500- or 300-um sieve; enteropneusts from Antarctica and off Oregon were decanted from sediment directly onto a 250-um sieve. Freshly collected worms were preserved in 95%-100% ethanol, and when multiple specimens were available, voucher specimens were relaxed in 7.5% magnesium chloride and fixed in 4%-10% formalin for morphological studies.

Molecular methods
Molecular laboratory methods followed Cannon et al. (2009). In addition to in-house sequencing on a CEQ8000 (Beckman-Coulter, Inc., Brea, CA), some PCR products were sent to Genewiz (South Plainfield, NJ). In the case of Rhabdopleura cf. normani from Bermuda, we were able to generate a full-length 18S rDNA sequence using publically available partial sequences (Genbank Accession numbers: U15664, JF900483, and JF900484) and 454 transcriptome data (Cannon and Halanych, unpubl.) prepared following the methods of Kocot et al. (2011). Assembled 454 contigs were formatted into a BLAST (Altschul et al., 1997) database, which was searched with a query composed of diverse hemichordate 18S sequences. Fragments with E-values below 1x10^30 were aligned to the partial 18S sequences as a backbone, and then assembled into a full-length 18S rDNA cofntig using the software package CLC Genomics Workbench ver. 5 (Aarhus, Denmark). As reported previously (Cannon et al. 2009), Stereobalanus canadensis (Enteropneusta, Harrimaniidae) has a highly divergent 18S rDNA sequence. Previously we had sequenced this acorn worm multiple times, but to further verify this sequence, genomic DNA was extracted from S. canadensis and sequenced via an Illumina MiSeq (San Diego, CA) at Auburn University. This work employed a 2x150-bp paired end run and the Nextera DNA sample kit. Scaffolds generated using the genome assembler Ray (Boisvert et al., 2010) were formatted into a BLAST database, and queried as above. A single 18S rDNA sequence was recovered, with 99.6% identity to our previously reported sequence; thus we have retained Stereobalanus canadensis in our analyses.

Taxon sampling and NCBI accession numbers for sequences used in phylogenetic analyses are given in Table 2. Representatives of each echinoderm class were used as outgroup taxa (Table 2). Sequences were aligned with MAFFT ver. 6.09b (Katoh et al., 2005) using the L-INS-i method, and uninformative sites were trimmed using the perl scripts Aliscore and Alicut (Misof and Misof, 2009). Models of evolution were selected under the Akaike Information Criterion (AIC) implemented by MrModelTest (Nylander, 2004). The best fitting model for both 16S and 18S rDNA, GTR+I+G, was used in maximum likelihood (ML) and Bayesian inference analyses (BI) of concatenated alignments. ML analyses were performed with RaxML version 7.3.9 (Stamatakis, 2006) using 5000 bootstrap replicates. Bayesian analyses were conducted using MrBayes ver. 3.2.0 (Ronquist and Huelsenbeck, 2003). Four independent BI analyses were run for each dataset for 5,000,000 generations with trees sampled every 100 generations using three heated and one cold chain. Plotting likelihood values versus generation number revealed that stationarity was reached after approximately 1,000,000 generations, and thus the first 25% of sampled trees was discarded as burn-in. Competing hypotheses of hemichordate phylogeny were evaluated using the SH-test (Shimodaira, 2002) as implemented in RAxML ver. 7.3.8 with the GTR+I+G model.