Data for: Environment and host genetics influence the biogeography of plant microbiome structure
Description
We collected Lemna and associated microbiomes from 34 populations in the northern and southern range of its distribution in the United States (Fig. 1a and Table S1): Ohio (OH, Cleveland, N = 8; Columbus, N = 5), New Hampshire (NH, N = 2), Massachusetts (MA, N = 2), Rhode Island (RI, N = 2), Louisiana (LA, N = 7), Georgia (GA, N = 4), and South Carolina (SC, N = 4). The field sampling was conducted during the fast-growing season of duckweeds during June–August 2022. In addition, we collected samples from the same two Massachusetts populations during the late growing season in October 2022 to confirm the negligible influence of temporal dynamics on duckweed microbiomes, relative to the other factors we investigated in this study. We also measured the pH, conductivity (EC), and total dissolved solids (TDS) of the aquatic environment at each population using an Ohaus ST20M-B meter (Ohaus Corporation, Parsippany, New Jersey). Additionally, we collected 100 mL surface water in sterile centrifuge tubes and sent to the Wetland Biochemistry Analytical Services at Louisiana State University for additional water chemistry analysis (total organic carbon, TOC; total nitrogen, TN; total phosphorus, TP; major and trace elements including Na, Ca, Mg, Fe, Si, Cu, Zn, Mn, Pb, Cd; Table S1). Duckweed microbiome DNA was sent to the Argonne National Laboratory for bacterial library preparation (16S rRNA V5–V6 region, 799f–1115r primer pair: AACMGGATTAGATACCCKG, AGGGTTGCGCTCGTTG) and sequencing using Illumina MiSeq (paired-end 250 bp). The PE reads were used for detecting bacterial amplicon sequence variants (ASVs) using the package DADA2 v1.20.0 in R v4.1.0. The PE reads were trimmed and quality filtered [truncLen = c(240, 230), trimLeft=c(10, 0), maxN = 0, truncQ = 2, maxEE = c(2,2)] and then used for unique sequence identification that took into account sequence errors. The PE reads were then end joined (minOverlap = 20, maxMismatch = 4) for ASV detection and chimera removal. The ASVs were assigned with taxonomic identification based on the SILVA reference database (132 release NR 99) implemented in DADA2. The ASVs were further filtered before conversion into a bacterial community matrix using the package phyloseq. First, we removed non-focal ASVs (Archaea, chloroplasts, and mitochondria). Second, we conducted rarefaction analysis using the package iNEXT to confirm that the sequencing effort was sufficient to capture duckweed bacterial richness (Fig. S1). We further normalized per-sample reads (median = 20,192 reads) by rarefying to 10,000 reads. Three populations that had fewer reads (one from OH: 9787 reads; two from GA: 5775 and 9484 reads, respectively) but plateaued in the rarefaction analysis were normalized to 10,000 reads. Lastly, we removed low-frequency ASVs (<0.001% of total observations). The final bacterial community matrix consisted of 4880 ASVs across the 36 samples from 34 different populations and was used for all downstream analyses.