Sequencing and analysis

Library concentrations were normalized and pooled at 5 µL volumes at 4 nM before denaturing with 5 µL of 0.2 N NaOH. Samples were centrifuged at 280 g for 1 min, and then incubated for 5 min at room temperature. A Hybridization Buffer (Illumina) was added to the denatured libraries to dilute to 4–6 pM and the internal standard, PhiX Control Kit V3 (Illumina), was added. Multiplexed libraries were sequenced on an Illumina MiSeq.

Bioinformatic analysis was conducted using QIIME2 version 2019.10³⁴. Samples were demultiplexed and trimmed to 280 bp to maintain a quality threshold of 30 or more (see SI Table ¹ for read number per mesocosm). Reads from technical replicates of each mesocosm were merged to increase sampling depth. Sequences were denoised to ASVs using the DADA2 pipeline within QIIME2 software³⁵ Then, alpha rarefaction was performed on data from both markers to estimate where saturation of metrics (e.g., observed feature) occurred. To generate the core metrics, COI sequences were sampled to a depth of 50,371 (retained 79.9% features in 100% of the samples). For 18S sequences the depth was 18,223 (retained 69.6% features in 100% of the samples). Rarefaction curves for both makers indicate that the number of observed features plateaued well below the sampling depths chosen (see Fig. ^S2). Sequences were classified using QIIME2 from the SILVA_132 18S database and the protozoan and invertebrate sequences from the Barcode of Life Database (BoLD) for 18S and COI libraries, respectively. Further identification of unassigned ASVs from the initial classification was performed using the NCBI Basic Local Alignment Search Tool (BLAST). Sequences were clustered into Amplicon Sequence Variants (ASVs) based on classification. Observed ASVs and Shannon Index α-diversity metrics as well as Bray–Curtis and Jaccard’s distance β-diversity metrics were calculated using QIIME2. Bray–Curtis dissimilarity was selected as a β-diversity metric because it is a quantitative measurement that accounts for relative abundance of each ASV. Jaccard’s distance was selected because it is a qualitative measurement that is based on the presence or absence of each ASV.