Dataset: NCBI accession numbers and related metadata for an SRA archive of the seagrass, Zostera marina

During the June peak of seasonal flowering in 2019, divers on SCUBA collected a single vegetative shoot at each node within each sampling grid for genetic analysis (n=15 ramets per grid, 45 per depth, 90 per site). At approximately alternating nodes (i.e., 5-7 of the 15 nodes in each grid, depending on the location of bare patches), divers also harvested all shoots from within a 25 × 25 centimeter (cm) quadrat (n = 15-21 per depth) to estimate shoot density and canopy characteristics. Samples were returned to the lab on ice for processing. The second and third leaves of each genetic sample were preserved in silica.

After the flowering season and once seeds had dropped (September 2019; von Staats et al. 2021), we returned to each sampling grid to collect dispersed seeds. Sediment cores (10 cm in diameter and 10 cm in depth) were taken at three locations within each grid at Curlew Beach and at four locations within each grid at the other three field sites (n=9 or 12 cores per depth). Sediment cores were bagged and kept cold (4 degrees Celsius (°C)) until they could be processed; each core was hand-sieved for intact seeds within 3 days of collection. We counted all intact seeds encountered and assessed viability by firmness (the squeeze test; Marion and Orth 2010). Seeds deemed viable were individually weighed and stored in microfuge tubes and frozen at -80°C until DNA extraction.

Leaf samples (2 to 4 milligrams (mg) of dried tissue from the middle third of the leaf) were ground with a Retsch mixer mill MM400; seeds were ground by hand with a micropestle after removing the seed coat. DNA extraction was done with the Omega Bio-Tek E-Z 96 Tissue DNA kit, and samples were stored at -20°C. We prepared one genomic library of 464 individuals (359 adults, 105 seeds) by following the ddRADseq protocol of Parchman et al. (2012) and sequenced the library using two lanes of Illumina Novaseq. We digested gDNA with two restriction enzymes, EcoRI and MseI, and ligated adaptors containing unique 8-10bp barcodes to the digested DNA of each individual. The products were then PCR-amplified in two independent reactions with standard Illumina primers. All amplicons were pooled and shipped to the University of Texas Genomic Sequencing and Analysis Facility, which used Blue Pippen Prep to isolate the 300-450bp fraction. This fraction was then single-read-sequenced (approximately 100bp) with both lanes of an Illumina Novaseq machine. We used custom scripts to demultiplex into sample-specific FASTQ- formatted files.