Table of Contents

• Coverage
• Dataset Description
  • Methods & Sampling
  • Data Processing Description
• Data Files
• Related Publications
• Parameters
• Instruments
• Deployments
• Project Information
• Funding

For most samples, plankton biomass for Pseudo-nitzschia DNA identification was collected by passing an average of 270 mL of surface seawater with a peristaltic pump across a 25 mm 5.0 mm polyester membrane filter (Sterlitech, Kent, WA, USA). Widths of some Pseudo-nitzschia spp. are < 5.0 mm (Lelong et al. 2012), but this size pore likely captured horizontally orientated cells and chains of cells, and was consistent with pore size used to examine toxicity. Filters were flash frozen in liquid nitrogen and stored at -80 °C until extraction. DNA was extracted using a modified version of the DNeasy Plant DNA extraction kit (Qiagen, Germantown, MD, USA) with an added bead beating step for 1 minute and QIA-Shredder column (Qiagen, Germantown, MD, USA) as reported in Chappell et al. 2019. Additionally, DNA was eluted in 30 µL with a second elution step of either 30 or 15 µL to maximize DNA yield. DNA was assessed for quality with a Nanodrop spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA) and quantified using a Qubit fluorometer (Invitrogen, Carlsbad, CA, USA) with the Broad Range dsDNA and High Sensitivity dsDNA kits (Thermo Fisher Scientific Inc., Waltham, MA, USA). DNA yields reported by the Qubit ranged from below the limit of detection to 26.5, with an average of 2.0 ng DNA / mL eluent. Long-Term Plankton Time Series (LTPTS) samples from October 2016 and March 2017 had an average of 300 mL surface seawater passed over a 25 mm 0.2 mm filter, were extracted following existing LTPTS methods of DNA extraction using the DNeasy Blood and Tissue Kit (Qiagen, Germantown, MD, USA) with an added bead beating step (Canesi and Rynearson 2016), and yielded average 0.9 ng DNA / mL eluent as measured by the Qubit. Net tow samples had 50 mL of concentrate was passed across a 0.22 µm pore size Sterivex filter unit (MilliporeSigma, Burlington, MA, USA), and were extracted with the same modified DNeasy Plant DNA extraction protocol as above, with 4x volumes of AP1 buffer and RNase A and beads added to the unit to account for the larger sample surface area, extraction occurring within the capped unit itself to maximize yield, and then the lysate removed with a sterile syringe and subsequent steps with adjusted volumes as appropriate. As expected, DNA yields were higher from the Sterivex units ranging from 2.4 – 54.0 ng DNA / mL eluent with an average of 13.7 ng DNA/ mL elution as measured by the Qubit. For the March 13, 2017 NBay samples, 125 mL of surface seawater was passed across a HV filter and extracted with the DNeasy Plant DNA extraction kit with scissors and no beads. As measured by the Qubit, the average DNA yield was 3.7 ng DNA / mL eluent. A negative control sample was prepared of a blank 25 mm 5.0 mm polyester membrane filter using extraction reagents which had no detectable DNA using the Qubit. There were two positive controls of mock communities comprised of two known Pseudo-nitzschia species from monocultures. The two Pseudo-nitzschia cultures were P. subcurvata collected from the Southern Ocean and P. pungens isolated from NBay (provided by J. Rines). One positive control was made by combining equal concentrations of extracted DNA with 1.0 ng DNA of each culture. The second positive control was created of equal cell abundance estimated to be captured onto the filters of the cultures prior to extraction. These negative and positive controls were prepared for sequencing and sequenced on the same plate as the other environmental samples.

The ITS1 has been targeted for amplification and analysis by ARISA previously for Pseudo-nitzschia identification in environmental samples (Hubbard, Rocap, and Armbrust 2008). A comparison of ITS1 appears to be much less conserved and is divergent enough across Pseudo-nitzschia that 41 different species can be identified using existing public sequencing data. The primers to target the ITS1 region of Pseudo-nitzschia used this existing forward primer sequence of the ITS1 region for eukaryotes: TCCGTAGGTGAACCTGCGG (White et al. 1990) and a custom reverse primer designed using 132 Pseudo-nitzschia ITS1 sequences from the NCBI nucleotide database (downloaded on 4/3/2019) from this nucleotide search: ((Pseudo-nitzschia[Organism]) AND internal transcribed spacer[Title]) NOT uncultured): CATCCACCGCTGAAAGTTGTAA. This reverse primer targets a conserved region in the 5.8S. All primer sequences are reported from 5’ – 3’. MiSeq adapter sequences were added to the beginning of the primer sequences for these full sequences used in this study: forward primer TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGTCCGTAGGTGAACCTGCGG and reverse primer GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGCATCCACCGCTGAAAGTTGTAA. When checking the specificity of these primers using the NCBI nt database, it became known that sequences beyond Pseudo-nitzschia would also be amplified in this study including other diatoms and dinoflagellates; however, the large number of sequencing reads recovered on the MiSeq platform would circumvent this non-specific characteristic of the primers.

The accession numbers of the sequences used in this primer design are reported in Table S2 of Sterling et al. (in prep), along with a summary of Pseudo-nitzschia species expected to amplify with these based on the in silico design. The expected ranges for PCR products were from 235 – 370 bp as the size of the ITS1 region differs for some Pseudo-nitzschia taxa. Primers (Integrated DNA Technologies, Coralville, IA, USA) were HPLC purified, resuspended in 1x Tris-Acetate-EDTA (TAE) buffer, and then working stocks created in diethylpyrocarbonate (DEPC)-treated H2O. About 4 ng of extracted DNA was used for each PCR reaction. If, according to the Qubit quantification, the DNA concentration was less than 2 ng mL-1 or below the limit of detection, it was then used as is, and just 2 mL was added to the PCR reaction. PCR reactions were set up on ice, in a 1x reaction in 25 mL total volume. Final primer concentration was 0.5 mM and polymerase was Phusion Hot Start High-Fidelity Master Mix (Thermo Fisher Scientific Inc., Waltham, MA, USA). There were two cycles with different annealing temperatures, the first with an annealing temperature specific to the loci-specific region and the second set of cycles with an annealing temperature that also takes the MiSeq adapter sequence into account (Canesi and Rynearson 2016). PCR conditions used were initial denaturation for 30 seconds at 98 °C, 15 cycles of the following: denaturation for 10 seconds at 98 °C, annealing for 30 seconds at 64.1 °C , extension for 30 seconds at 72 °C, and 15 cycles with the same conditions except a higher annealing temperature of 72 °C , and then a final extension for 10 minutes at 72 °C , and a holding temperature of 10 °C until stored in the -20 °C freezer. PCR products were visualized on a 1% agarose gel before submission to the URI Genomics and Sequencing Center (Kington, RI, USA) where library preparation and sequencing were performed on a 2x300 bp MiSeq run (Illumina, Inc., San Diego, CA, USA). There were 193 environmental samples were sequenced, along with two positive controls of Pseudo-nitzschia DNA from cultures and one negative control, for a total of 196 samples using two sets of MiSeq indices on the same sequencing plate. It was deemed appropriate to multiplex this plate as estimated read depth to recover Pseudo-nitzschia sequences was predicted to be lower than usual.

The columns are the library_ID as described in DATASET 01: the identifying sample number that connects the row of environmental data to the corresponding plankton biomass filter that was sequenced for the Pseudo-nitzschia species assemblages. The Sequence Sample ID connect to the sample_title, library_ID and file names in NCBI’s Short Read Archive (SRA) under this related Bioproject #PRJNA690940.

The rows are the Sequence_of_ASV as described in DATASET 02: the DNA sequence of the internal transcribed spacer (ITS) 1 region used to identify the Pseudo-nitzschia [NCBI:txid41953] species. Only ASVs shown pass the threshold of accounting for > 1% relative abundance in a sample.

The matrix is filled in indicating whether or not that specific ASV sequence was present in that corresponding sample with 1 = Present and 0 = Absent.

Problem report: Sample #AS424 had no ASVs belonging to Pseudo-nitzschia sp. and was removed.

A custom bioinformatics pipeline was utilized. CutAdapt (Martin 2011) was used to trim Illumina MiSeq adapters and primer sequences. Primer sequences were trimmed from both ends of sequences, with the reverse complement of the other primer trimmed the end of the sequences. If reads did not have the ITS1 primer sequence, they were discarded. Reads needed to be one base pair (bp) or longer to continue in the pipeline. Trimmed sequences were inputted into DADA2 (v. 1.16)  to determine amplicon sequence variants (ASVs; Callahan et al. 2016). ASVs were retained at that level, with some potentially having as few differences as one bp to each other, for the subsequent analysis. ASVs were identified as Pseudo-nitzschia taxa using a curated database from NCBI sequences (Table S2 in Sterling et al. in prep) which used to design primers to assign taxonomy for ITS1 ASVs trimmed of the primer sequences using the scikit-learn naïve Bayes machine learning classifier (Pedregosa et al. 2011) at default settings in QIIME2 (Bolyen et al. 2019). The scikit-learn naïve Bayes machine learning classifier identified 97 ASVs as Pseudo-nitzschia at the species level. Three of these ASVs belonged to P. subcurvata from the positive control mock community and were removed from analysis. All of the 6,503 ASVs recovered from the 192 non-control samples from the sequencing effort were run through a megablast search using BLAST+ version 2.9 with the nucleaotide (nt) database downloaded on October 4, 2020. There were 540 ASVs which had a known Pseudo-nitzschia taxa, including clones, vouchers, and environmental samples, as its top megablast hit. In addition to the 97 ASVs identified as a specific Pseudo-nitzschia species from the QIIME2 pipeline, there were 115 ASVs identified as a Pseudo-nitzschia taxa with greater than 75% query coverage were manually examined. It was determined by judgement call that the 11 ASVs which were identified as P. pungens PC50 were likely Cylindrotheca instead and the 85 ASVs which were closest related to P. delicatissima KJ22-0.2-69 environmental clone was most closely related to known Nitzschia isolate sequence from subsequent BLAST searches. This left 19 ASVs of interest, with 9 of them have >98% query coverage and >98% identity with known Pseudo-nitzschia sequences so were referred to as the specific Pseudo-nitzschia species and 10 ASVs were identified as the genus with identifiers of similar groups of ASVs to each other. These genus level ASVs have < 96% identity to existing sequences in the database. In total, there were 113 ASVs from the 192 samples that appeared to be of reliable Pseudo-nitzschia origin. Sample #AS424 had none of the 113 ASVs and was removed. Read counts were transformed into relative abundance out of total Pseudo-nitzschia taxa reads. If an ASV accounted for < 1% relative abundance in a sample, then it was considered “not present” or absent to avoid potentially spurious results. This removed 60 ASVs which only occurred in < 1% of reads in samples. The remaining 53 ASVs were used in the analysis in a presence/absence matrix to avoid potential problems from inflating read numbers with cell counts. This threshold retained 46 of the 97 scikit-learn classifier identified ASVs, and seven of the ASVs added by the megablast curation. Of the seven ASVs added from megablast results, three ASVs where in a group together at the genus level, and around 95% identity with known P. americana sequences. The other megablast added ASVs were very closely related to P. cuspidata and P. calliantha.

BCO-DMO Processing Notes:
- data were submitted in file "DATA03_ASVTable_Sterling_NBay.csv".
- added conventional header with dataset name, PI name, version date
- columns and rows were flipped to allow better viewing (would be extremely wide if not pivoted)

NSF Award Abstract:

Non-technical Description
The University of Rhode Island (URI) will establish the Consortium for Coastal Ecology Assessment, Innovation, and Modeling (C-AIM) to coordinate research, education, and workforce development across Rhode Island (RI) in coastal marine science and ecology. C-AIM addresses fundamental research questions using observations, computational methods, and technology development applied to Narraganset Bay (NB), the largest estuary in New England and home to important ecosystem services including fisheries, recreation, and tourism. The research will improve understanding of the microorganisms in NB, develop new models to predict pollution and harmful algal bloom events in NB, build new sensors for nutrients and pollutants, and provide data and tools for stakeholders in the state. Observational capabilities will be coordinated in an open platform for researchers across RI; it will provide real-time physical, chemical, and biological observations ? including live streaming to mobile devices. C-AIM will also establish the RI STEAM (STEM + Art) Imaging Consortium to foster collaboration between artists, designers, engineers, and scientists. Research internships will be offered to undergraduate students throughout the state and seed funding for research projects will be competitively awarded to Primarily Undergraduate Institution partners.

Technical Description
C-AIM will employ observations and modeling to assess interactions between organisms and ecosystem function in NB and investigate ecological responses to environmental events, such as hypoxia and algal blooms. Observations of the circulation, biogeochemistry, and ecosystem will be made using existing and new instrument platforms. The Bay Observatory ? a network of observational platforms around NB - will be networked to trigger enhanced water sampling and sensing during specific environmental events, such as hypoxic conditions or phytoplankton blooms. Biogeochemical, ecological, and coastal circulation models will be integrated and coupled to focus on eutrophication and pollutant loading. Data and models will be integrated on multiple scales, from individual organisms and trophic interactions to food-web responses, and from turbulence to the regional ocean circulation. New sensing technologies for nutrients and pollutants will be developed, including affordable, micro-fluidic (Lab-on-a-Chip) devices with antifouling capabilities. The results will be synthesized and communicated to stakeholders.