Ribosomal rRNA identification and relative abundance
\nFrom the high-quality paired-end Illumina sequencing reads, 16S rRNA gene fragments were identified using Meta-RNA (Huang et al., 2009; v.H3; -e 1e-10). Putative rRNA fragments and associated mate pairs from each sample were processed through EMIRGE (Miller et al., 2011, 2013); emirge_amplicon.py; -l 113 -i 163 -s 33 -a 32 \u2014phred33) to generate full-length sequences using the SILVA (Quast et al., 2012) SSURef111 reference database (https://github.com/csmiller/EMIRGE). Reconstructed 16S rRNA genes were assigned taxonomy using mothur (v1.34.4) by first aligning the sequences to the SILVA SSURef123 database (align.seqs; flip=T), removing sequences that failed to align, if necessary (remove.seqs), and classifying the sequences (classify.seqs; cutoff=80, iters=1000).
Crustal fluids were collected from the single horizon at U1382A and from the shallow, middle and deep horizons in U1383C (Edwards et al., 2012) using a mobile pumping system designed for microbial sampling from CORK fluid delivery lines as described in Meyer et al. (2016) and Cowen et al. (2012). Deployed with the ROV system, mobile pumping system connectors are attached to the CORK wellhead via an umbilical to the hydrological zone of interest within the aquifer.\u00a0 In 2012, 12 l of each fluid sample were filtered on to a 0.22 \u03bcm Sterivex-GP filter (Merck Millipore, Billerica, MA, USA) as described in Meyer et al. (2016). In 2014, 12 l of each sample was filtered in situ and immediately fixed with RNALater (Thermo Fisher Scientific, Waltham, MA, USA), as described previously (Akerman et al., 2013). After sampling in 2012, a battery-powered GeoMICROBE sled was left at each CORK for time series autonomous sampling of the fluid delivery lines (Cowen et al., 2012). For each filter sample, ~10 l of fluid were filtered in situ and immediately fixed with RNALater.\u00a0 Sleds were deployed in April 2012 and recovered in April 2014 with samples collected. Upon sled recovery, filters were transferred to fresh RNALater and stored at \u221280 \u00b0C, while all bag samples were stored at 4 \u00b0C (Cowen et al., 2012). Deep bottom water was sampled in 2012 and 2014 via a CTD at 100 m above the seafloor and filtered in the same manner as the crustal fluids onto Sterivex filters. Total microbial biomass in fluids was enumerated with DAPI (4\u2032,6\u2032-diamidino-2-phenylindole; Sigma-Aldrich, St Louis, MO, USA) and epifluorescent microscopy (Porter and Feig, 1980).\u00a0
BCO-DMO Processing Notes:
\n-\u00a0added conventional header with dataset name, PI name, version date
\n- modified parameter names to conform with BCO-DMO naming conventions
\n
\u00a0