Introduction

Corals are keystone species in reef habitats, serving a multitude of roles in maintaining a healthy ecosystem1. In performing these roles, they are reliant on the symbioses that exist between the host and its microbial constituents, including, but not limited to bacteria and Symbiodiniaceae2,3. These symbiotic relationships are crucial for the healthy functioning of the coral holobiont, with bacteria involved in nutrient cycling, metabolism and immune system functioning4,5, while Symbiodiniaceae provide a significant proportion of their hosts nutritional requirements through photoautotrophy6. Importantly, these symbionts can also help promote adaptation to environmental change7,8,9,10,11. Given this, an understanding of the symbiotic microbial communities and how they may change is an important consideration in efforts to understand how corals are expected to adapt as climate change advances. However, without baseline work describing the current community structure and microbial compositions, determining the magnitude and direction of any changes is impossible.

Corals display a variety of life-history strategies that can influence microbial community composition12, with coral species, life-history and traits (e.g., competitive, weedy, stress-tolerant, opportunistic or generalist) all influencing microbial community assemblage13,14,15. Competitive species have fast growth rates and reproduce via broadcast spawning, whereas species classified as weedy tend to be brooders with smaller colony sizes and shorter generation times. Stress tolerant species have slow growth rates tending to reproduce by broadcast spawning and are highly fecund, with species described as generalist or opportunistic having moderate growth rates and a degree of overlap with competitive, weedy and stress-tolerant life histories, for a complete description of coral life history traits see Darling et al. 201216. Additionally, in terms of how corals shape their microbial community, they can be divided into two groups (i.e., microbial conformers or regulators)12. Conformers have a microbiome that changes in response to the environment, while regulators are able to maintain stable microbiomes across environments. Similarly, some microbial taxa show specialized versus generalist behaviors. Members of the Symbiodiniaceae genus, Durusdinium, specifically D. trenchii are considered more heat-tolerant and can confer an advantage to their host in elevated temperatures17,18,19, and some bacteria are implicated in the suppression of bacterial pathogens and play a role in maintaining a healthy microbiome, while others are involved in sulphur and nutrient cycling processes20. Here, we investigate the bacterial and Symbiombiodiniaceae communities of four coral species, Diploastrea heliopora, Porites lutea, Pachyseris speciosa and Pocillopora acuta, collected around the Malaysian state of Sabah (Borneo). Specifically, we determine whether (1) these communities show local spatial patterns and host-specificity, and (2) a ‘core microbiome’ is present within and amongst species. The concept of a ‘core microbiome’ was initially explored in humans where it was defined as a group of microbial taxa shared by all, or most humans21. The shared nature of these microbes led to suggestions of an important role in the maintenance of host functioning22, similar ideas have been applied in natural ecosystems where it is hypothesised that this core has roles in the maintenance of ecological functioning9,23,24. Here we apply this concept in an effort to determine whether a ‘core microbiome’ exists between the four coral host species collected in this work.

We hypothesise that the bacterial and Symbiodiniaceae communities will differ by host species, a consequence of their different life history strategies e.g., D. heliopora and P. lutea demonstrate stress tolerant strategies and traits, while P. speciosa is a generalist16,25 and P. acuta is classed as an opportunistic species26. Additionally, but to a lesser extent in comparison to host species, we expect that site specific environmental conditions will influence microbial community composition27.

This study seeks to characterise the microbial diversity associated with reef-building corals in an area that remains understudied but contains highly diverse and rich coral reef communities. Much of the world’s coral reef biodiversity resides in the global south, yet much of the research capacity is focused on the much less diverse and arguably less healthy reefs of the global north28. If coral reefs are to persist into the future it is critical that we work on understanding these megadiverse marine ecosystems and their essential microbial constituents, especially those at the epicenter of global marine biodiversity, for example the Coral Triangle and its surrounding regions. It is here that the evolutionary novelty required to adapt to a rapidly changing climate likely resides.

Methods

Sample collection

Corals from four sites in Sabah, Malaysia were collected between February and March 2020. Tissue samples measuring approximately 2 cm2 were collected from each of the following species Diploastrea heliopora, Pachyseris speciosa, Pocillopora acuta, and Porites lutea. Collections were made from reefs in Kota Kinabalu, Labuan, Lankayan, and Mataking, of these four sites Mataking is the only one located within the Coral Triangle biodiversity hotspot (Fig. 1), at each location 20 individual samples from each species were taken. All samples were collected from depths between 5 and 10 m and appeared visibly healthy with no apparent signs of bleaching or disease. Upon collection, samples were placed in individual, sterile Whirl–Pak bags and placed in 100% ethanol within 2 h, they were then stored at − 20 °C until transferred to − 80 °C prior to DNA extraction.

Fig. 1
figure 1

Location of sampling sites in Sabah, Malaysia. Figure was created using The Global Multi-Resolution Topography tool (https://gmrt.org/).

Full size image

DNA extraction and library preparation

DNA extraction was performed with a Qiagen DNeasy Blood and Tissue Kit and followed all manufacturer’s instructions. PCR amplification of the bacterial 16S rRNA gene V4 region was performed using the 515F and 806R primers (515F: 5′- GTGYCAGCMGCCGCGGTAA -3′; 806R: 5′- GGACTACNVGGGTWTCTAAT-3′) and modified to include Illumina adaptors and unique barcodes following Caporaso et al.29. To reduce Plastid and mitochondrial amplification, peptide nucleic acids (PNAs) were added (mPNA: GGC AAG TGT TCT TCG GA; pPNA: GGC TCA ACC CTG GAC AG) (PNAGENE, Daejeon, South Korea). We followed the thermocycling conditions and protocol described in30. Briefly, PCR was performed in 25 µl volumes containing 1 μl of undiluted template, 0.1 μl of KAPA 3G Enzyme (Kapa Biosystems, Inc, Wilmington, MA, USA), 0.75 μl of each primer at 10 μM, 2.5 μl of mPNA and 2.5 μl pPNA at 50 μM, 1.5 μl of 1.5 mg/ml BSA, 12.5 μl KAPA PCR buffer and water to 25 µl. PCR cycling protocol was 94 °C for 180 s, followed by 35 cycles of 94 °C for 45 s, 75 °C for 10 s, 50 °C for 60 s and 72 °C for 90 s, with a final extension at 72 °C for 10 min. Negative controls were included to identify any potential contamination.

For Symbiodiniaceae, the nuclear ribosomal internal transcribed spacer 2 (ITS2) region was amplified following the two-step protocol described in31. Briefly, the ITS2 region was amplified using the SYM_VAR primer pair (SYM_VAR_5.8S2: 5′–GAA TTG CAG AAC TCC GTG AAC C–3′, SYM_VAR_REV: 5′–CGG GTT CWC TTG TYT GAC TTC ATG C–3′)32 in 12.5 µl reaction volumes containing 1 µl of sample DNA diluted 1:10 with PCR grade water, 0.5 µl of forward and reverse primers each at 10 μM, 6.25 µl of 2 × KAPA HiFi HotStart ReadyMix and water to make a final volume of 12.5 µl. PCR cycling conditions consisted of an initial step of 95 °C for 3 min followed by 25 cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s, with a final extension step at 72 °C for 5 min. PCR products were then cleaned using AMPure XP magnetic beads (Agencourt). A reduced cycle PCR was used to attach unique barcodes and Illumina adaptors 33, reactions were performed in 12.5 µl volumes containing 2 µL of amplicon from the previous round, 0.5 µl of each uniquely indexed primer at 10 μM, 6.25 µL of 2 × KAPA HiFi HotStart ReadyMix, and water to a total volume of 12.5 µl. PCR cycling conditions were 95 °C for 3 min for initial denaturation, followed by 8 cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s, with a final extension step at 72 °C for 5 min. PCR negatives were included again to identify any possible contamination.

Prior to pooling, PCR products were cleaned and normalized using SequalPrep™ Normalization Plates (Thermo Fisher Scientific, Waltham, MA, USA). Samples were pooled into two libraries (bacteria or Symbiodiniaceae) and each library was sequenced independently on the Illumina MiSeq platform (V3 chemistry, 300 bp paired-end reads, 30% Phi-X spike), with sequencing performed by Macrogen, Inc.

Sequencing and data processing

All bioinformatic steps, unless mentioned otherwise, were performed in R v4.3.134. For 16S rRNA gene amplicons, cutadapt v4.4 implemented in Python v3.11.4 was used to remove primers from demultiplexed sequences35,36. Quality profiles of reads were checked and reads filtered and trimmed using the DADA2 package v1.28.037. D. heliopora and P. speciosa forward and reverse reads were trimmed to 250 and 150 base pairs respectively. Filtering was performed with a maximum expected error setting of 2 for both forward and reverse reads. P. acuta and P. lutea forward and reverse reads were trimmed to 180 and 120 base pairs, respectively, with a maximum expected error setting of 2 and 5 for forward and reverse reads. All other settings remained at their defaults. Error rates were then estimated and paired dereplicated reads were used infer amplicon sequence variants (ASVs). Any chimeras were identified and removed along with contaminants using control “blank” samples with the prevalence method implemented in the decontam package v1.20.038. Taxonomic assignment was performed, to species level where possible, based on the SILVA v138.1 16S rRNA gene database39. ASVs that were assigned as “mitochondria” or “chloroplast” were removed. ASVs were further filtered to retain only those that were found in at least 5% of all samples40. Other than three P. speciosa samples (Kota Kinabalu: n = 2, Mataking: n = 1) all samples passed quality control and were used in downstream analysis.

Analysis of ITS2 Symbiodiniaceae samples was performed using SymPortal v0.3.2441. All steps were performed using SymPortal’s default settings. Sequence filtering and quality control were performed using MOTHUR, BLASTn was used identify Symbiodiniaceae, and minimum entropy decomposition (MED) nodes identified42,43. Defining intragenomic sequence variants (DIVs) and downstream type profiles of Symbiodiniaceae were then inferred within the SymPortal framework to assess community composition. Downstream analysis was performed for both DIVs and type profiles separately.

Community analysis

Symbiodiniaceae and bacterial communities were visualized on separate non-metric multidimensional scaling (nMDS) plots using the Bray–Curtis metric with ellipses indicating 95% confidence intervals. Permutational multivariate ANOVA (PERMANOVA) was performed with Bray–Curtis distances with host species and site as predictors using vegan v2.6.444. The Shannon diversity index for each sample was calculated and grouped by host species and site. Relative abundance bar plots by host species and site were constructed for samples to show distribution of bacterial phyla, and Symbiodiniaceae genera and type profiles. Only the top 20 ASVs and type profiles from each species were used for these plots. A plot of the majority sequence for type profiles was constructed to simplify the type profile relative abundance. This majority sequence is the sequence that makes up the largest abundance of each type profile, as identified by SymPortal. Individual sample relative abundance plots were also constructed and relative abundance was also plotted as heatmaps, grouped by host species and site.

The core microbiome was arbitrarily defined as bacterial ASVs present in at least 50% of individuals with a detection threshold of 1%45,46,47,48. The core_members function from the microbiome package v1.22.0 was used to extract these core microbes for all samples and for each host species49. These sequences were searched using BLASTn to infer their individual characteristics and presence in other studies.

Results

Bacteria

After quality control 3,171,579 bacterial reads were retained for analysis. Ordinations indicate that bacterial communities can be differentiated by host species, with clustering by species observed (Fig. 2a). This community differentiation is further supported by our PERMANOVA results which indicated significant differences between species (PREMANOVA R2 = 0.22, F = 32.55 p = 0.001). Plots of Shannon diversity indicate that the bacterial communities in P. speciosa and D. heliopora have the highest and second highest diversities in all but one location, Mataking. With bacterial diversity across all four species at Mataking appearing broadly similar (Fig. 3). In three of the four sites bacterial diversity is lowest in P. acuta (Fig. 3a). Our PERMANOVA tests indicate significant structuring in bacterial communities by site (PREMANOVA R2 = 0.08, F = 11.15 p = 0.001), with this structure supported by our ordinations (Fig. 4a). The bacterial phylum Proteobacteria is largely dominant across all host species, and is especially prevalent in P. lutea and P. acuta (SI Fig. 1–3). Bacteroidota has the highest abundance in P. speciosa (mean relative abundance 20.9 ± SD 0.10 and is largely absent in P. lutea and P. acuta (mean relative abundance 0.45 ± SD 0.002 and 3.41 ± SD 0.04 respectively) (SI Fig. 1).

Fig. 2
figure 2

Non-metric multidimensional scaling (nMDS) plots of Bray–Curtis distances for (a) bacterial ASVs and (b) Symbiodiniaceae sequences. Ellipses indicate 95% confidence intervals.

Full size image
Fig. 3
figure 3

Shannon diversity plots of (a) bacterial communities and (b) Symbiodiniaceae profiles.

Full size image
Fig. 4
figure 4

Non-metric multidimensional scaling (nMDS) plots of (a) bacterial ASVs and (b) Symbiodiniaceae sequences by each host species. Based upon the Bray–Curtis metric with ellipses indicating 95% confidence intervals.

Full size image

Using the criteria defined above to identify a core microbiome (ASVs present in at least 50% of individuals with a detection threshold of 1%) D. heliopora had a core of 13 bacterial ASVs belonging to eight families and P. speciosa had 48 bacterial ASVs from 32 families. We found four bacterial ASVs from three bacterial families in the P. acuta samples, with the P. lutea core microbiome containing 16 bacterial ASVs from six families (Fig. 6, SI Fig. 4). Out of these ASVs, six were shared in the core microbiome of three host species while none were shared among all four host species’ core microbiome. The six ASVs shared between the three species belong to four families: Alteromonadaceae, Cyanobiaceae, Vibrionaceae, and Pseudoalteromonadaceae. Using the criteria described to define the core microbiome, no ASVs in the core microbiome are shared. However, 22 ASVs are found to be present across all four hosts.

Symbiodiniaceae

After quality control, 9,375,672 Symbiodiniaceae reads were retained for analysis. Ordinations indicate that Symbiodiniaceae communities can be differentiated by host species, (Fig. 2b). This community differentiation is further supported by our PERMANOVA results which indicated significant differences between species (PREMANOVA R2 = 0.12, F = 15.04 p = 0.001). For Symbiodiniaceae communities, P. speciosa shows variable diversity across all sites, of the four species it has the highest diversity at Kota Kinabalu and Labuan, while it shows the lowest diversity at Lankayan and Mataking. Symbiodiniaceae diversity in P. acuta is similar across Kota Kinabalu, Labuan and Lankayan, but of all four species it has the highest diversity in Mataking. While Symbiodiniaceae diversity in P. lutea and D. heliopora is broadly comparable across all four sites, and generally lower in comparison to the other species examined (Fig. 3b). However, it should be noted that rDNA copy number in Cladocopium can be five times higher than Durusdinium, therefore Symbiodiniaceae diversity metrics should be treated with a degree of caution, especially in mixed endosymbiont communities where both Cladocopium and Durusdinium are present. Nevertheless, these metrics can still give a broad over view of diversity provided these limitations are considered50. Symbiodiniaceae communities can be significantly differentiated by site (PREMANOVA R2 = 0.05, F = 6.30 p = 0.001). This is further supported by the ordinations, while there is evidence of clustering, the separation of communities is less well defined in comparison to microbial communities (Fig. 4a, b). Across all four species we identified 20 Symbiodiniaceae type profiles with all Symbiodiniaceae belonging to the genera, Cladocopium or Durusdinium (Fig. 5b, SI Fig. 5, 6). Each of the four host species are dominated by different type profiles, in D. heliopora, three of the four sites are dominated by type profile C3u/C3/C115, while Mataking is dominated by C3 and has C40, and type C40/C3 that is not found in the other locations. P. speciosa samples are largely composed of type C27 and D1, whereas Mataking has a reduced abundance of C27, a high abundance of C21 and C27/C15h types. P. acuta samples from three of the four sample sites are primarily composed of type profiles D1, or D1b, with Mataking differing markedly from other sites again, P. acuta from Mataking is dominated by C1d/C1, and C1d, both of which are absent in the other locations. In comparison to the other three host species, P. lutea, has a much more uniform Symbiodiniaceae type profile that is dominated by type C15 across all sites.

Fig. 5
figure 5

Bar plots showing relative abundance of (a) family of top 20 bacterial taxa and (b) the majority sequence of the top 20 Symbiodiniaceae profiles for each host species.

Full size image
Fig. 6
figure 6

Heatmaps (a) showing relative abundance of all amplicon sequence variants (ASVs) found in the core microbiomes of each individual coral host species (b) ASVs identified as belonging to the core microbiome using the predefined criteria (occurring in 50% of all individuals with a detection threshold of 1%). KK Kota Kinabalu, LA Labuan, LY Lankayan, MT Mataking.

Full size image

Discussion

This is the first examination and characterization of coral associated microbes to be performed around the Malaysian state of Sabah, on the island of Borneo. Collections spanned a distance of nearly 1000 km and include part of the Coral Triangle diversity hotspot as well as the South China Sea. Unsurprisingly, and similar to other studies performed throughout the Southeast Asian region26,27,51 we show that bacterial and Symbiodiniaceae communities from the four collected host species of coral are structured by, and can be differentiated by geographic location and species.

Bacterial community patterns

Life history characteristics of the coral host can be useful when describing the relationship between host and its associated microbial communities, for example, Diploastrea heliopora and Porites lutea are considered stress tolerant species52, Pocillopora acuta is described as an opportunistic colonizer, while Pachyseris speciosa is considered a generalist53,54. Broadly in line with these characteristics, and as previously reported26 we find that the more stress tolerant species tend to have less diverse microbial communities, whereas the more opportunistic and generalist species have a more diverse microbial communities which should allow them to persist in a wider range of conditions. This pattern is evident in bacterial communities with diversity highest in the generalist, P. speciosa. But, contrary to expectations the opportunistic colonizer P. acuta has the lowest bacterial diversity. Further work is needed to confirm this, but it is possible that P. acuta is a microbiome regulator and as such is able to select its microbial constituents. Adding weight to this suggestion, the low microbial diversity is indicative of a microbiome that has been tailored to host specific requirements through microbial winnowing55 with similar, low bacterial diversity reported in this species previously56,57. These species-specific differences are supported by our ordinations that show well defined clusters indicating differences in the hosted bacterial communities between coral species. Diploastrea heliopora shows the lowest degree of differentiation in bacterial community structure, this is consistent with work performed in Peninsular Malaysia48 and could be a consequence of stress tolerant species requiring fewer symbiotic partners for survival26,58. A similar, but less pronounced pattern is seen in P. lutea, another species described as stress tolerant. Bacterial community structure is more defined in Pocillopora acuta and Pachyseris speciosa, described as opportunistic or as a generalist respectively. It is likely that the life history characteristics of P. acuta and P. speciosa allows these species to take advantage of disturbance or exist in a wide range of habitats, to do this, it is probable they acquire bacteria that are prevalent in the local environment, hence we see clear differentiation in community structure between sampling locations. In fact, a number of studies exploring the microbiome of P. acuta and how it responds to new environments when transplanted indicate that it is strongly influenced by and quickly conforms to local conditions59,60,61.

The bacterial communities in all four coral species are dominated by Proteobacteria, and to a lesser extent Bacteroidota, Cyanobacteria, and Firmicutes. These general associations are well documented in corals throughout the world11,14,62,63,64,65. The bacterial family Colwelliaceae observed in P. speciosa and P. lutea has been implicated in buffering the negative effects of photooxidative stress linked to bleaching66, this family would be a good candidate for future research attempting to determine how microbes can contribute to coral stress adaptation. This association fits well with the generalist and stress tolerant life histories these corals have, and may in part be responsible for their ability to persist in stressful conditions.

Symbiodiniaceae community patterns

Coral-associated symbidiniaceae communities typically vary by species and geographic location26,67. The Porites lutea colonies examined here are dominated by Symbiodiniaceae type C15 and its variants, this C15 association is found throughout Southeast Asia52,67,68,69 and more generally throughout the world70. Similar to other work performed in the South China Sea71 we find the Diploastrea heliopora host is dominated by type C3u/C3/C115 at three of the four sites, whereas this type is absent from Mataking and of the four sites type C40/C3 is only present here. At all sites, and as previously reported from the Malay Peninsula26,58, type C27 is most prevalent in Pachyseris speciosa, with type C21 only found at Mataking. Pocillopora acuta is dominated by types D1 and D1b in three of the sites, whereas Mataking is dominated by types C1d and C1d/C1. The Symbiodiniaceae community composition of P. lutea shows little variation across all four sites, while the other three species have similar Symbiodiniaceae communities at Kota Kinabalu, Labuan and Lankayan, whereas Symbiodiniaceae communities differ considerably in P. speciosa, D. heliopora and P. acuta collected from Mataking in comparison to the other sample sites (Fig. 5b). This difference is particularly striking in P. acuta where a large shift from Durusdinium to Cladocopium is seen. Further work is needed to confirm this and with the data we have, we can only speculate why the samples collected from Mataking are frequently different. Unlike the other collection sites, Mataking is in close proximity to the deep waters of the Celebes Sea, whereas the other three sites are in shallower coastal water associated with the Sunda Shelf. Shallower water does heat up faster and corals found here could experience warmer waters and have shifted their Symbiodiniaceae communities in response to the more stress tolerant Durusdinium, members of this genus, particularly D. trenchii have been shown to confer an advantage in warmer waters17,18,71. Similar phenomena has previously been reported in Pocillopora spp from the eastern tropical pacific and has been suggested as a mechanism that may confer increased resilience to ocean warming11. Whereas the deeper water of the Celebes Sea could buffer and prevent heat stress in corals from Mataking, consequently we see a different Symbiodiniaceae community, one that is adapted to the different conditions experienced here. This site is also the only one within the Coral Triangle biodiversity hotspot, an area where coral biodiversity reaches its maxima72. Work shows that the composition of the coral community in the immediate area can influence the endosymbionts found in a host73. Given the elevated coral diversity of the Coral Triangle region, this could, in part also be responsible for the different Symbiodiniaceae communities seen at Mataking and could play a role in structuring Symbiodiniaceae communities in hosts that acquire their symbionts from the environment, rather than through vertical transmission. However, it should be noted that despite having significantly lower coral host diversity, previous work suggests the Caribbean actually has more Symbiodiniaceae species than the Indo-Pacific74. Although, this may change as high throughput DNA sequencing approaches are increasingly applied to examine Symbiodiniaceae communities and diversity throughout the Coral Triangle and other hotspots of marine biodiversity.

Both D. heliopora and P. speciosa have more mixed Symbiodiniaceae profiles at Kota Kinabalu in comparison to P. acuata and P. lutea, mixed communities have been implicated as advantageous in certain environmental conditions such as elevated temperatures or low light environments75,76. As the capital of Sabah, it is not unreasonable to suggest that the corals found at Kota Kinabalu are subject to the increased anthropogenic stresses associated with a human population that is larger than that at other locations in Sabah (e.g., terrestrial runoff). Further work is needed to confirm this, but, this mixed profile maybe a response to the increased stresses unique to Kota Kinabulu.

Core microbiome

Six core microbial ASVs are shared between three of the four corals—ASV2, ASV67, ASV1, ASV3, ASV6, and ASV7. These ASVs have been identified in other work examining coral microbial communities, this ubiquity suggests a role in promoting host health. ASV2 (Alteromonas spp) and ASV67 (Synechococcus CC9902 spp) are shared between P. speciosa, P. acuta, and P. lutea, while the remaining four are shared among D. heliopora, P. speciosa, and P. lutea.

Pseudoalteromonas (ASV 7) and Vibrio (ASVs 1 & 3) have been observed associated with numerous coral hosts, suggesting important roles of these bacteria in maintaining microbiome functioning48,77,78. ASVs 6 (Algicola bacteriolytica) & 7 (Pseudoalteromonas spp) are involved in dimethylsulfoniopropionate (DMSP) metabolism, DMSP metabolizers are part of the Beneficial Microorganisms for Corals (BMC) group and have been identified as having important roles in nutrient cycling and preventing bacterial pathogen outbreak, helping to maintain the hosts microbiome10. ASV2 has been implicated as having potential roles in nitrogen fixation and could offer pathways of alternative fixation during periods of stress (e.g., thermal bleaching) and thus increase coral host resilience79. ASV67 has been recovered from coral tissue where it is thought to have roles in nutrient uptake80,81,82.

Studies such as this are increasingly common place throughout the world, however, they remain comparatively rare in the global epicenter of marine biodiversity. Given this, studies such as the one performed here are valuable as they provide an initial foundation upon which we can build and develop more hypothesis driven research. For example, future work will more specifically examine differences in environmental conditions and coral community composition between sampling sites around the island of Borneo to determine why differences in microbial community composition exist. Additionally, as coral reef restoration activities become increasingly popular, it will become more important that we understand the diversity contained within an ecosystem, for example, if we do not know what is there already, how will we determine if restoration is successful? Additionally, a more complete understanding of the microbial constituents that are present in marine ecosystems could help restoration practitioners determine and prioritize the most suitable candidate sites for restoration.