Dataset: Acropora cervicornis 16s amplicon rDNA sequencing data with associated data on tank exposure to white band disease and survival outcomes.

Coral fragments (n=550) from 50 Acropora cervicornis genotypes were collected from Coral Restoration Foundation's in-situ nursery in Tavernier, Florida Keys in June 2021 (permit CRF-2021-001). One fragment from 18 genotypes was immediately sampled (day 0); three polyps per fragment were preserved in DNA/RNA Shield (Zymo Research) at -20°C. Remaining fragments (10 per genotype) were distributed across 10 recirculating 18-liter tanks at Florida Keys Marine Laboratory (50 fragments/tank), lesioned with a WaterPik, and exposed to either healthy or diseased tissue homogenate doses (50 mL, 5 tanks each). Homogenates were prepared from five healthy and five diseased nursery-collected fragments and normalized by optical density. Fragments were monitored twice daily for white band disease signs. Three polyps were sampled from each fragment on day 3 and day 7 (or when diseased), preserved in DNA/RNA Shield at -20°C for <2 months. Diseased fragments were removed after sampling. Total samples sequenced: 234 (18 day 0, 206 tank samples across 46 genotypes, 10 homogenate doses).

Genomic DNA was extracted using GenElute Bacterial Genomic DNA kit (Sigma-Aldrich). The V3-V4 region of bacterial 16S rRNA gene was amplified using primers S-D-Bact-0341b-S-17 (5'-CCTACGGGNGGCWGCAG-3') and S-D-Bact-0785-a-A-21 (5'-GACTACHVGGGTATCTAATCC-3') (Herlemann et al. 2011, Klindworth et al. 2013). First PCR (23 μL): 1 μL DNA, 1.25 μL each primer (10 μM), 12.5 μL Phusion Mix (2X, Thermo Fisher), 7 μL water; cycling: 98°C 1 min, 28 cycles (98°C 30s, 63°C 30s, 72°C 30s), 72°C 5 min. Products cleaned with ZR-96 DNA Clean-Up kit (Zymo Research). Second PCR added Illumina Nextera XT indexes (25 μL): 5 μL purified product, 2.5 μL each index, 12.5 μL Phusion Mix, 2.5 μL water; cycling: 98°C 1 min, 12 cycles (98°C 30s, 55°C 30s, 72°C 30s), 72°C 5 min. Indexed products were normalized (SequalPrep, Thermo Fisher), pooled, and concentrated (DNA Clean & Concentrator, Zymo Research). Libraries sequenced on two Illumina MiSeq runs (2×300 bp, v3 chemistry).

Raw reads processed using DADA2 v1.28.0 (Callahan et al. 2016) in R v4.4.1: filterAndTrim (trimLeft=25, truncLen=c(250,230), maxEE=c(2,2)); dereplication; error rate estimation on first 40 samples (err=NULL, selfConsist=TRUE); sample inference (pool=TRUE); paired read merging; ASV table generation; chimera removal (removeBimeraDenovo, consensus method). Taxonomic classification used BLCA v3.0 (Gao et al. 2017) with NCBI 16S database (80% confidence, accessed 2/5/2024); ASVs unclassified below class reclassified using SILVA SSU r138 (Quast et al. 2013, 50% threshold). ASV sequences aligned with DECIPHER v2.26.0 (Wright 2016); neighbor-joining tree constructed with phangorn v2.11.1 (Schliep 2011). Data filtered to remove non-bacteria, chloroplast, mitochondria, cyanobacteria, and samples with <1000 reads. For differential abundance: removed ASVs in <20% samples and ASVs absent from day 0 and homogenate doses. Final filtered dataset: 254 ASVs across 211 samples; unfiltered (5,716 ASVs) used for visualization and alpha diversity. Normalized using TMM with effective library size (ELib-TMM, Robinson & Oshlack 2010) in edgeR v3.42.4 (Robinson et al. 2010); transformed to log₂ CPM + 0.5. Data managed in phyloseq v1.44.0 (McMurdie & Holmes 2013).

Alpha diversity (Shannon, Simpson) calculated on rarefied data (1029 reads, seed 68748) using microbiome v1.22.0 (Lahti & Shetty 2019); tested with linear mixed-effects models (fixed: time, exposure, outcome; random: genotype, tank). Beta diversity: Bray-Curtis dissimilarity (Bray & Curtis 1957) visualized by NMDS; PERMANOVA tested effects of time, exposure, outcome, genotype, tank (10,000 permutations, seed 68748). Disease resistance correlations: Kendall's tau between day 0 abundances (ASV, genus, family) and resistance scores (Vollmer et al. 2023); FDR-corrected. Differential abundance: weighted linear mixed-effects models using variancePartition v1.30.2 with dream (Hoffman & Roussos 2021, Hoffman & Schadt 2016); weights from voomWithDreamWeights (Law et al. 2014); fixed effects: time, exposure, outcome; random: genotype, tank. Post-hoc contrasts using emmeans v1.8.9: (1) transplant effect: day 0 vs. mean(day 3+7 healthy); (2) exposure: healthy- vs. disease-exposed; (3) outcome: diseased vs. healthy; (4) temporal response on diseased corals: day 0 vs. 3 (early), day 3 vs. 7 (late); (5) outcome within disease-exposed: diseased vs. healthy at peak timepoint (day 3 for early, day 7 for late responders); (6) doses: diseased vs. healthy (linear models). All P-values FDR-corrected (Benjamini & Hochberg 1995, α=0.05). Putative pathogens required: positive disease outcome association, late response (day 3-7 increase), higher abundance on diseased vs. healthy in disease-exposed tanks (day 7), and enrichment in diseased dose. Community composition visualized using Fantaxtic v0.2.0 (Teunisse 2022).