Metagenomic analysis
Microbial DNA from the coral mucus collected on the 0.22-µm Sterivex was extracted using a modified Macherey-Nagel protocol using NucleoSpin column for purification. DNA was stored at - 20°C until quantification with Qubit (Thermo Fisher Scientific) (Dinsdale et al. 2008b). The Swift kit 2S plus (Swift Biosciences) was used for library preparation since it provides good results from small amounts of input DNA, characteristic of microbial samples collected from the surface of the host (Doane et al. 2017; Cavalcanti et al. 2018). All samples were sequenced by the Dinsdale lab on Illumina MiSeq at San Diego State University. The sequenced DNA was analyzed for quality control using PrinSeq (Schmieder and Edwards 2011) before annotation. The metagenomes were annotated through MG-RAST (Meyer et al. 2019), using the RefSeq database for taxonomic annotations and the SEED database for functional annotations.
For the taxonomic composition, the metagenomes were filtered in MG-RAST to include only Bacteria at the level of genera. For the functional composition, the metagenomes were filtered in MG-RAST for Stress Response, Nitrogen Metabolism, and Sulfur Metabolism. We selected these three broad functional gene groups (SEED subsystem level 1) because they have the greatest relevance for corals under heat stress (Wegley et al. 2007; Thurber et al. 2009; Littman et al. 2011; Raina et al. 2013; Nguyen-Kim et al. 2015; Cardini et al. 2016; Pogoreutz et al. 2017; Wang et al. 2018; Sun et al. 2022). The number of sequence hits for each microbial taxon or function is represented as the relative abundance by calculating the proportion of sequence hits for that parameter over the total number of sequences annotated for that metagenome. Metagenomes were compared using proportional abundance, which is preferred to rarefaction (McMurdie and Holmes 2014; Quince et al. 2017; Luz Calle 2019). Bacteria accounted for approximately 99 % of the annotation; therefore, we are only analyzing bacterial taxa and gene functions in this study. Metagenomes were sequenced from a subset of the coral nubbins used in the experiment (n = 8), representing four different coral colonies (n = 2 colonies per reef zone) that had their mucus sampled before and after the experimental treatments (n = 16 metagenomes total). These metagenomes were compared with the ones from the same colonies collected in situ (Lima et al. 2022).