Dataset: RNA-Seq of Lake Erie microbial community

Water samples were collected in the western basin of Lake Erie during the summer and fall of 2017, 2018, and 2019.  In 2017, water was collected approximately weekly from NOAA station WE2 in conjunction with the NOAA Great Lakes Environmental Research Lab (GLERL) harmful algal bloom monitoring program.  During August and October 2017, lake water was also collected by Environment and Climate Change Canada’s monitoring program.  In 2018 and 2019, lake water was collected at several stages of bloom development (pre-bloom, early bloom, late bloom, and post bloom).  In 2018, lake water was collected at NOAA’s monitoring stations WE2 and WE12 and at the drinking water intake for the City of Toledo (TWI).  During summer 2019, the goal was to sample lake waters containing high bloom biomass as predicted by the NOAA HAB forecast model and HAB tracker bulletins (Wynne et al. 2013).  Sampling sites were chosen based on the presence of surface scums comprised of dense cyanobacterial colonies (i.e., “bloom chase” sites).

For all sites, a depth-integrated water sample was collected in acid-washed carboys.  Water samples were collected from the NOAA stations using a peristaltic pump.  The pump hose was moved down the water column from the surface to 1 meter above the bottom.  For the TWI, bloom chase, and Environment Canada cruise sites, a depth-integrated sample was collected by pooling water collected at 1 m intervals from surface to 1 meter above the lake bottom using a Niskin (Environment Canada) or Van Dorn (TWI and bloom chase sites) bottle.  Integrated water samples were stored in carboys in an outdoor aquaculture tank until the start of the bottle experiments the following morning.  The water temperature in the aquaculture tank was controlled using copper piping attached to a NESLAB RTE refrigerated water bath (Thermo Scientific, Newington, NH) and maintained at the lake temperature measured at the time of sample collection.  During the Environment Canada cruises, bottles and carboys were stored in a plexiglass tank continuously circulated with fresh lake water.

Subsamples for supporting water quality analyses were taken from each carboy.  Upon arrival in the laboratory at the University of Michigan, a subsample of whole (unfiltered) water was taken for analysis of total phosphorus.  During 2017, pH of the water from each site was obtained from NOAA monitoring buoys.  For samples collected in 2018-2019, pH of the whole water was measured upon arrival in the laboratory.   Subsamples of the whole water were filtered through a 0.22 μm polyethersulfone (PES) filter for subsequent analysis of total dissolved phosphorus (TDP), soluble reactive phosphorous, nitrate and ammonium, dissolved organic carbon (DOC), and chromophoric and fluorescent dissolved organic matter (CDOM and FDOM, respectively).  DOC samples were preserved by addition of 6N trace metal grade hydrochloric acid to pH 3.   TDP, SRP, DOC, CDOM and FDOM were stored in the dark at 4 °C until analysis.  Nitrate and ammonium samples were stored at -20 °C until analysis at GLERL.

To measure microbial community composition, DNA was extracted from the filters collected from each bottle at the beginning (T=0) of each experiment using a Qiagen Dneasy Blood and Tissue Kit with QIAshredder columns (QIAGEN, MD, USA).  The extraction protocol is included as a supplemental file.  For absolute quantification of sequence data, genomic DNA from Thermus thermophilus strain DSM 7039 was added to the samples after the cell lysis step of the extraction as an internal standard. Thermus thermophilus DNA was obtained from the American Type Culture Collection (ATCC; product number BAA-163D-5).  The internal standard was added as ~ 1 % of DNA yield, which was estimated based on an empirically determined relationship between total mass of chlorophyll a on the filter and DNA yield.  DNA yields were measured with Quant-iT Picogreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA).  The true percentage of the internal standard was 0.72 ± 0.37 % of total DNA yield, on average.

The V4 region of the bacterial 16S rRNA gene was amplified using a dual-indexed primer set (Kozich et al. 2013), and amplicon sequencing was performed using Illumina MiSeq V2 500 cycle chemistry (Illumina cat# MS102-2003) at the University of Michigan Microbial Community Analysis Core following Schloss and Bishop (2019).  Forward and reverse reads were quality screened to remove sequences below 250 bases and trimmed to Q20 using the BBDuk tool in BBTools (Bushnell 2018). Following trimming, overlapping forward and reverse reads were assembled into contigs, aligned, screened for chimeras, and clustered into operational taxonomic units (OTUs) using MOTHUR v. 1.43.0, following the protocol as of February 2020 (Kozich et al. 2013).  OTU clustering was performed using a 97 % similarity cutoff with the OptiClust algorithm (Westcott and Schloss 2017).  Contigs were aligned with the align.seqs function in MOTHUR, and taxonomy was assigned using the Wang method (Wang et al., 2007).  The Silva v. 138 SSU database (Pruesse et al. 2007) was used as the reference to align and classify contigs.  The absolute abundance of each OTU per volume of lake water was estimated from the recovery of the internal standard as described in (Lin et al. 2019). All OTUs classified as Genus Thermus were removed from downstream analysis.  The raw sequence data are available in NCBI under BioProject PRJNA646259.