Data for: Environment and host genetics influence the biogeography of plant microbiome structure

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.本研究从美国浮萍属（Lemna）分布范围的南北区域共34个种群中采集浮萍样本及其附着微生物群（microbiomes）（图1a、表S1）：俄亥俄州（OH，克利夫兰，样本量N=8；哥伦布，N=5）、新罕布什尔州（NH，N=2）、马萨诸塞州（MA，N=2）、罗德岛州（RI，N=2）、路易斯安那州（LA，N=7）、佐治亚州（GA，N=4）以及南卡罗来纳州（SC，N=4）。野外采样于2022年6-8月的浮萍快速生长季开展。此外，为确认相较于本研究考察的其他因素，时间动态对浮萍微生物群的影响可忽略不计，我们于2022年10月的生长季后期，从马萨诸塞州的2个原采样种群中补充采集了样本。我们使用Ohaus ST20M-B型检测仪（奥豪斯公司，新泽西州帕西帕尼）测定了每个采样点水生环境的pH值、电导率（Electrical Conductivity, EC）及总溶解固体（Total Dissolved Solids, TDS）。另外，我们用无菌离心管采集100 mL地表水，送至路易斯安那州立大学湿地生物化学分析服务中心开展额外的水化学分析，包括总有机碳（Total Organic Carbon, TOC）、总氮（Total Nitrogen, TN）、总磷（Total Phosphorus, TP），以及Na、Ca、Mg、Fe、Si、Cu、Zn、Mn、Pb、Cd等常量与微量元素（详见表S1）。 浮萍微生物组DNA被送往阿尔贡国家实验室（Argonne National Laboratory）进行细菌文库构建：针对16S rRNA V5-V6高变区，采用引物对799f–1115r（序列为AACMGGATTAGATACCCKG、AGGGTTGCGCTCGTTG），并使用Illumina MiSeq平台进行双端250 bp测序。我们采用R v4.1.0环境下的DADA2 v1.20.0包，基于双端reads识别细菌扩增子序列变异体（Amplicon Sequence Variant, ASV）。首先对双端reads进行修剪与质量过滤，参数设置为truncLen = c(240, 230)、trimLeft=c(10, 0)、maxN = 0、truncQ = 2、maxEE = c(2,2)；随后利用过滤后的reads识别考虑序列误差的唯一序列。接着将双端reads进行末端拼接（最小重叠区minOverlap=20，最大错配数maxMismatch=4），以完成ASV检测与嵌合体移除。基于DADA2中集成的SILVA参考数据库（132版NR 99），对ASVs进行分类学注释。在使用phyloseq包将ASVs转换为细菌群落矩阵前，需先对其进行进一步筛选：第一步移除非目标ASVs（古菌、叶绿体及线粒体序列）；第二步使用iNEXT包开展稀疏化分析，确认测序量足以覆盖浮萍细菌群落的丰富度（详见图S1）。随后通过抽平至10000条reads的方式实现样本间测序量标准化，所有样本的原始reads中位数为20192条。对于3个reads数不足的种群（俄亥俄州1个：9787条reads；佐治亚州2个：分别为5775和9484条reads），尽管其稀疏化曲线已趋于平稳，仍统一抽平至10000条reads。最后移除占总观测值<0.001%的低频率ASVs。最终得到的细菌群落矩阵包含来自34个种群的36个样本中的4880个ASVs，用于后续所有分析。