This dataset includes RNA-seq differential expression results of Pfam domains within diatoms from an iron incubation experiment using upwelled waters sampled in the California Current System (CCS) during the PUPCYCLE II cruise with Chief Scientist Adrian Marchetti. PUPCYCLE II (Phytoplankton response to the UPwelling CYCLE) took place onboard the R/V Sally Ride from May 29th to June 10th, 2023. Freshly upwelled seawater for the incubation experiment was collected within the northern CCS and placed into designated cubitainers. Three cubitainers were immediately harvested for the initial timepoint (T0). The remaining twenty-seven cubitainers were assigned treatments: nine were unamended (Ctrl), nine were amended with 5 nM FeCl2 (+Fe), and nine were amended with 200 nM Desferrioxamine B, a strong iron chelator (DFB). Three cubitainers from each treatment were harvested for each of the three subsequent timepoints: 48 hours (T1), 168 hours (T2), and 264 hours (T3) after incubation. Approximately 2.5 to 4 liters of each sample was harvested for RNA-seq of the natural phytoplankton community, or metatranscriptomics. The dataset provides differential expression results of Pfam domains within diatoms (and which genera they map to), specifically comparing expression levels between +Fe and DFB at T1 and T2. These data include domains that significantly differ in expression between treatments in diatoms, and results were used to compare how strict phototrophs like diatoms transcriptionally respond to iron bioavailability compared to mixotrophs.

Table of Contents

• Coverage
• Dataset Description
  • Methods & Sampling
  • Data Processing Description
  • BCO-DMO Curation Notes
• Data Files
• Related Publications
• Related Datasets
• Parameters
• Instruments
• Deployments
• Project Information
• Funding

This dataset is one of many generated from an iron incubation experiment conducted as part of the PUPCYCLE II cruise in May and June 2023. The different analyses are listed here, with links to other datasets from this study in the Related Datasets section.

Analyses include:

• Physiology (dataset 994607)
  • Dissolved inorganic nutrients (nitrate, phosphate, silicic acid)
  • Size-fractionated chlorophyll a
  • Size-fractionated isotope uptake rates (nitrate and DIC)
  • Flow cytometry
  • FlowCAM analysis
  • FLP experiments for mixotroph cell-specific grazing rates
• RNA taxonomy data for normalized reads (dataset 1003525)
• Differential expression for +Fe versus DFB treatments at T1 and T2: Pfam
  • Mixotroph Pfam (dataset 1004659)
  • Diatom Pfam (this dataset 1004674)
• Differential expression for +Fe versus DFB treatments at T1 and T2 for mixotrophs and diatoms: KEGG gene (dataset 1004689)
• Differential expression for +Fe versus DFB treatments at T1 and T2 for mixotrophs and diatoms: KEGG pathways (this dataset 1004704)
• Dissolved iron concentrations (dataset 1004007)

See Related Datasets section below for links to above mentioned datasets.

Data collection took place on the R/V Sally Ride from May 29th to June 10th, 2023. To simulate upwelling conditions under different iron treatments, an onboard incubation experiment was conducted. Seawater for the incubation experiment was collected within the northern CCS, off the southern coast of Oregon at 43°02'42.7"N, 124°33'07.2"W. The collected seawater was deemed as freshly upwelled water and taken from a depth of 55 meters – which corresponds to a depth slightly below the euphotic zone receiving less than 1% irradiance – using trace-metal clean techniques on May 30th, 2023, 13:30 GMT. The seawater was pumped and homogenized using trace metal clean techniques, then transferred into a total of thirty 20 L low-density polyethylene cubitainers. Three cubitainers were immediately harvested for the initial timepoint (T0). The remaining twenty-seven cubitainers were assigned treatments: nine were unamended (Ctrl), nine were amended with 5 nM FeCl₂ (+Fe), and nine were amended with 200 nM desferrioxamine B, a strong iron chelator that inhibits dissolved iron uptake (DFB). The cubitainers were placed in an on-deck incubator covered with two layers of neutral density screening to simulate 26% of incident irradiance. The incubator was supplied with flow-through surface seawater to maintain ambient surface temperature. Three cubitainers from each treatment were harvested for each of the three subsequent timepoints: 48 hours (T1), 168 hours (T2), and 264 hours (T3) after incubation.

RNA Collection and Metatranscriptomics. Cubitainers were harvested for RNA by collecting and filtering approximately 2.5 to 4 L of each sample onto 0.8 μm Pall Supor filters (142 mm) using a peristaltic pump; samples were then immediately flash frozen in liquid nitrogen and stored at -80°C. RNA was extracted using the RNAqueous-4PCR kit, according to the manufacturer's instructions with a few modifications – filters were cut up due to their large size, 200 μL of glass beads were added, and 3 mL of lysis buffer was added. RNA samples were sent to GENEWIZ for library preparation and sequencing with PolyA tail selection. Sequencing was performed on an Illumina HiSeq 4000 with a 2x150 bp configuration. GENEWIZ provided raw paired-end read sequences for each sample. 

Raw reads were trimmed using Trim Galore v0.6.10 (Krueger, 2015), which uses Cutadapt (Martin, 2011), and quality control was assessed through FastQC (Andrews, 2010).  A de novo metatranscriptome assembly was conducted using rnaSPAdes v3.15.5 for individual assemblies (Bushmanova et al., 2019) and CD-HIT v4.8.1 for a grand assembly (Li & Godzik, 2006). TransDecoder v5.7.1 was used to determine the most likely open reading frame (ORF) for each contig (Haas, 2023). ORFs were annotated using the Marine Functional Eukaryotic Reference Taxa (MarFERReT) database v1.1.1 (e-value < 1e^-06) (Groussman et al., 2023) for NCBI taxonomies (Federhen, 2011), PR² taxonomies (Guillou et al., 2012) and Pfam 34.0 functions (Mistry et al., 2021). To remove any potential bacterial contamination, ORFs were also annotated against the PhyloDB v1076 database using EUKulele (e-value < 1e^-06) (Krinos et al., 2020). Any ORF with a higher bitscore for a bacteria or virus compared to its MarFERReT annotation was removed (resulting in < 3% of ORFs removed). To maximize functional annotation coverage, ORFs were annotated using eggNOG-mapper v2.1.12 (Cantalapiedra et al., 2021) to obtain annotations from the Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa, 2000). Trimmed samples were then aligned to the grand assembly using Salmon v1.10.3 (Patro et al., 2017) and exported into a comprehensive counts table using tximport (Soneson et al., 2016). Reads not mapping to a protist group as defined by MarFERReT annotations were removed before downstream analysis. 

DESeq2 was used to obtain normalized counts and conduct differential expression analyses (Love et al., 2014). Analyses of Pfam protein domains were performed on reads mapping to the lineage Bacillariophyta, hereafter referred to as “diatoms,” as defined by MarFERReT (Groussman et al., 2023). Reads not mapping to the taxonomic group being analyzed were removed before normalization and statistical testing. Reads were then summed by annotation ID (Pfam accession number) before entering the DESeq2 analysis. If a contig was annotated with multiple IDs, its reads were evenly distributed among the mapped annotations.

Differential expression analysis was conducted using the DESeq2 DESeq() function. This workflow estimates size factors using the median-of-ratios method, estimates dispersion, fits a negative binomial generalized linear model, and performs hypothesis testing using the Wald test. P values were adjusted for multiple testing using the Benjamini–Hochberg procedure to control the false discovery rate. Analyses focused on comparisons between +Fe and DFB treatments at T1 and T2. A Pfam protein domain was considered differentially expressed when the absolute value of its log2 fold change (L2FC) was greater than 1 and its adjusted P value (p-adj) was less than 0.05. Normalized counts for each Pfam protein domain were also obtained to determine the relative contributions of each diatom genus to each annotation ID.

Differential expression analyses for Pfam protein domains were performed for reads mapping to the lineage Bacillariophyta as defined by MarFERReT (henceforth referred to as diatoms). Reads not mapping to the taxa group being analyzed were removed prior to normalization and statistical testing. Reads were summed based on annotation ID (Pfam accession number) prior to entering DESeq2 analysis; if a contig annotated to multiple IDs, its reads were evenly distributed between the mapped annotations. A protein domain was considered differentially expressed if the absolute value of log2FoldChange (L2FC) > 1 and the p-adjusted value (p-adj) < 0.05

The metatranscriptomic pipeline code used for this project is publicly available at https://github.com/emilyspeciale/Speciale-Metatranscriptomics. 

Software: Trim Galore v0.6.10; FastQC v0.12.1; rnaSPAdes v3.15.5; CD-HIT v4.8.1; TransDecoder v5.7.1; EUKulele v2.0.9; eggNOG-mapper v2.1.12; Salmon v1.10.3;  tximport v1.3.9; DESeq2 v1.42.1.

Databases/reference resources: MarFERReT v1.1.1; NCBI Taxonomy; PR2; Pfam 34.0; PhyloDB v1076; Kyoto Encyclopedia of Genes and Genomes (KEGG); Mixoplankton Database (MDB).

• 1. Raw reads were trimmed with Trim Galore v0.6.10 (Krueger, 2015)
• 2. Read quality was assessed with FastQC v0.12.1 (Andrews, 2010)
• 3. De novo metatranscriptome assemblies were generated using rnaSPAdes v3.15.5 (Bushmanova et al., 2019) for individual assemblies and CD-HIT v4.8.1 (Li & Godzik, 2006) for the grand assembly
• 4. ORFs were identified with TransDecoder v5.7.1 (Haas, 2023). 
• 5. Taxonomy and functional annotation via Marine Functional Eukaryotic Reference Taxa (MarFERReT) database v1.1.1 (Groussman et al., 2023); includes NCBI taxonomies (Federhen, 2011), PR2 taxonomies (Guillou et al., 2012), and Pfam 34.0 functions (Mistry et al., 2021). 
• 6. PhyloDB v1076 filtering using EUKulele v2.0.9 (Krinos et al., 2020). 
• 7. eggNOG-mapper v2.1.12 (Cantalapiedra et al., 2021) / KEGG annotation (Kanehisa, 2000)
• 8. Reads were quantified with Salmon v1.10.3 (Patro et al., 2017) 
• 9. Reads were imported with tximport v1.3.9 (Soneson et al., 2016)
• 10. Taxonomic filtering where reads outside the target taxonomic group were removed
• 11. Annotation-ID read summation
• 12. DESeq2 v1.42.1 (Love et al., 2014) was used to compare +Fe versus DFB at T1 and T2

 

CURATION ACTIONS PERFORMED ON DATA
- Loaded source file "BCO_DMO_PUPCYCLEII_DiaPfamDE.csv" into the BCO-DMO system converting missing data identifier NA to empty cell
- Added latitude and longitude of sampling station
- Set data types to string for first 6 columns, and all other columns formatted as numeric
- Columns with p-value and padj were set to scientific notation to match how they appeared in the original file
- Reordered diatom genera to be in alphabetical order
- Exported final file as "1004674_v1_diatom_pfam.csv"

CURATION ACTIONS PERFORMED ON METADATA
- Changed in-line citations to have format of (author, year) replacing numbered references.
- Corrected first sentence of Abstract to say diatoms rather than mixotrophs
- Corrected minor typos and punctuation

ISSUES POTENTIALLY IMPACTING REUSE
- N/A

Coverage: California Upwelling Zone

NSF Award Abstract:
Upwelling zones are hotspots of photosynthesis that are very dynamic in space and time. Microscopic algae, known as phytoplankton, bloom when deep, nutrient-rich waters are upwelled into sunlit surface layers of the ocean, providing nourishment that supports productive food webs and draws down carbon dioxide (CO2) from the atmosphere to the deep ocean. Photosynthetic microbes in these regions must constantly adapt to changes in their chemical and physical environments. For example, subsurface populations respond to changes in light as they approach the surface. When upwelled waters move offshore, cells sink out of the illuminated zone, establishing seed populations that remain inactive until the next upwelling event. This process is called the upwelling conveyor belt cycle (UCBC). How phytoplankton respond to these changes in environmental conditions and how they may influence their nutrient requirements remains unknown. With future ocean changes predicted to alter seawater chemistry, including ocean acidification and decreased iron availability, some phytoplankton groups may be more vulnerable than others. Accompanying educational activities provide learning experiences to enhance understanding and awareness of marine microbes. The development of a research hub at UNC aims to provide infrastructure and support for scientists and students conducting research on environmental genomics. A laboratory component for an upper-level undergraduate course focused on marine phytoplankton is being developed. Educational outreach activities to broader communities include creation of a lesson plan on phytoplankton in upwelling zones and a virtual research cruise experience for middle-school students, as well as a hands-on lab activity for a local museum focused on marine phytoplankton and the important roles they play in shaping our planet.

The project examines how phytoplankton respond at the molecular and physiological level to the different UCBC stages, which seed populations (i.e., surface versus subsurface) contribute most to phytoplankton blooms during upwelling events of varying intensity, how phytoplankton elemental compositions are altered throughout UCBC stages, and how future predicted ocean conditions will affect the phytoplankton responses to UCBC conditions. This project contains both laboratory and fieldwork. In the laboratory, phytoplankton isolates recently obtained from upwelling regions are exposed to simulated UCBC conditions to examine changes in gene expression, growth and photosynthetic characteristics and elemental composition. Cultures are subjected to both current and future ocean conditions, including reduced iron availability and higher CO2. In the field, research cruises within upwelling regions study the dynamics of natural phytoplankton communities (both surface and subsurface) experiencing upwelling and relaxation and within simulated upwelling incubation experiments. Knowledge of how phytoplankton are affected by UCBC conditions at an integrated molecular, physiological and elemental level under both current and future scenarios is imperative for the proper conservation and management of these critically important ecosystems.