Reshuffling of the Coral Microbiome during Dormancy

As expected, there were fewer taxa in the active coral microbiome than in the present coral microbiome. This pattern was particularly evident in rarefied richness during and after dormancy (Fig. 2A, Table 2; active/present microbiomes, <0.05 for all diversity measures). Interestingly, for active microbiomes, we observed a significant decrease in alpha diversity as corals went into dormancy; it remained low as corals were in dormancy and then began to increase as corals exited dormancy (dormancy timing in D⁰ or richness: P = 0.004; D¹ or exponentiated Shannon diversity: P = 0.002; active/present microbiomes, <0.05; Fig. 2A and B, Table 2). However, in the present microbiome, there were no significant differences in diversity between before and during dormancy, but similarly, we observed an increase after dormancy based on Tukey honestly significant difference (HSD) tests (in D⁰ and D¹, Fig. 2A and B). Hill number D² or the inverse Simpson index, the diversity measure influenced by dominance, significantly increased only after corals came out of quiescence in both the active and present microbiomes (dormancy timing: P = 0.002; Fig. 2C).

Dispersion (beta diversity) was similar for both the active and present taxa (Fig. 3A and 3B, Table 2) and was consistent across the periods surrounding dormancy. However, the time point before corals went into quiescence (T3, 11 December 2020) showed significantly lower variability than that of all the other time periods, based on a Tukey HSD post hoc test (P < 0.05), and this was consistent in both the present and active microbiomes (Fig. 3A and B, Table 2; beta dispersion).

Astrangia poculata microbial community composition shifted significantly as corals went into quiescence and did not return to the same community after corals emerged from quiescence (Fig. 3c and D), suggesting a reshuffling of the microbiome that persisted even 2 months after corals were out of quiescence. Based on the permutational multivariate analysis of variance (PERMANOVA) analysis, we found significant effects of time (R² = 0.05, P < 0.001), dormancy (before, during, and after: R² = 0.08, P < 0.001), sample type (water/coral: R² = 0.15, P < 0.001), and active/present microbiomes (DNA/cDNA: R² = 0.02, P < 0.001).

Both the active (cDNA) and present (DNA) microbiomes changed in similar ways in nonmetric multidimensional scaling (NMDS) space (Fig. 3C and D), and we did not observe significant differences in microbial community composition within a time point between the active and present communities (Table S1; pairwise adonis results).

In both the active and present microbiomes, communities shifted markedly between before quiescence and during/after quiescence (time points 1 to 3 were different from time points 5 to 9) (Fig. 3C and D). The microbial community associated with the first time point for corals in quiescence (T4) was not significantly different from the time points before quiescence (T1 to T3) and from the time point 1 month later (T5), but it differed from all future time points (T6 to T9) (Fig. 3C and D; Table S1). Interestingly, postquiescent active microbiomes (time points 7, 8, and 9) did not significantly differ from corals in quiescence (time points 5 and 6) (Fig. 3C). The present microbiomes showed the same pattern, except that time point 9 (2 months after quiescence) was significantly different from the rest of the time points (Fig. 3D).

Seawater microbial community composition also changed over time significantly (R² = 0.72, P = 0.001), and there were significant differences in active/present seawater microbiomes (R² = 0.12, P = 0.001). The seawater microbiome was also consistently different from the coral microbiomes (Fig. 3C and D).

A total of 61 ASVs were identified to change significantly during the phases of quiescence in the active microbiome (Fig. S2). Several taxa that were higher in abundance before corals went into quiescence began to wane at the beginning of quiescence (time points 4 and 5). These taxa included Endozoicomonas, Arcobacter, two groups of Rickettsiales, and Pseudoalteromonas (Fig. 4A to F). During quiescence, the UBA10353 marine group showed a marked increase that lasted throughout the quiescent period (time points 4 to 6, Fig. 4G). In late quiescence and as corals began to emerge (time points 5 to 9), several taxa were enriched, including those in the orders Nitrosococcales (Fig. 4K and L, Cm1-21 and MSB-1D1) and Rhizobiales (Fig. 4M and N), and the genus Magnetospira (Fig. 4O; for a full list, see Fig. S2).

Enrichment in the total present microbiome followed a similar pattern at the order level; however, it included more ASVs that shifted in abundance and presence (a total of 126, Fig. S2). In particular, bacteria in orders Flavobacteriales, Chitinophagales, Cellvibrionales, and Sphingomonadales were higher as corals emerged from quiescence and during quiescence, compared to before corals went into quiescence. Additionally, an ASV of “Candidatus Nitrosopumilus,” a taxon frequently associated with A. poculata (15, 24), was enriched during and after corals came out of quiescence.

Some of these taxonomic changes are likely temperature or environment driven, as the same taxa in the water column shifted similarly in abundance (e.g., Synechococcus, Fig. S2). However, most of the significant taxon shifts in the coral were not observed in the water column (Fig. S2).

Across all time points, there were no taxa that were consistently present in the active microbiome of corals (100% of all samples), and only 5 ASVs were consistently present in 80% of samples across all time points (Bacteroidea, UBA4486, Terasakiellaceae, Endozoicomonas). Otherwise, core taxa changed by time point and varied between 4 and 8 ASVs within a time point, all of which were identified as significantly changing across dormancy time period in the corncob results (Fig. S3).

In the present microbiome, nine taxa were present in 80% of the samples (across all time periods). These taxa include those also found in the active microbiome and Persicirhabdus, Pirellulaceae, and Rubripirellula. Within a time point, core ASVs varied from 6 to 17 and included many that that changed significantly in the corncob results (Fig. S3).

We consistently observed two archaeal ASVs in the present core microbiome and in 65% of all active microbiomes which were associated with “Candidatus Nitrosopumilus” (Fig. S4). Only one of these ASVs was present in the water, and only in predormancy time points (Fig. S3). The other Nitrosopumilus ASVs we observed on the corals were in low relative abundances or not detectable in any seawater samples.

Of the 156 ASVs associated with significant shifts in the present and active microbiomes based on the corncob results, 106 taxa were assigned hypothesized functions from the FAPROTAX database. Based on the taxa that were identified as significantly enriched by the corncob results and their assignment of function with FAPROTAX, we found a reduction in the number of ASVs associated with photoautotrophy and photoheterotrophy on corals (in the cDNA and DNA) as coral went into and emerged from dormancy (Fig. S5A and B). We also observed a decline in the ASVs associated with intracellular parasites, nitrate reduction, sulfur/sulfite respiration, and methylotrophy in the coral active and present microbiomes as corals went into dormancy. ASVs associated with nitrogen fixation, nitrification, and dark sulfur oxidation increased after quiescence in the coral active microbiome.

Quiescence was associated with a streamlining of the coral’s microbiome. This was particularly evident in the active microbiome, as alpha diversity declined when corals entered dormancy. The loss of diversity is likely associated with the shedding of taxa associated with the dormancy period. Lowered alpha diversity is consistently a characteristic observed in dormant host-microbiome interactions, as diapausing parasitoid wasps (8) and ground squirrels also show a reduction in diversity of their dormant microbiomes (25).

As corals emerged from dormancy, we observed an increase in diversity (Hill numbers D⁰, D¹, D²) in both the active and present microbiome. We expect that one of the drivers of this increase in alpha diversity postquiescence is the increase in feeding by the corals (11) and thus greater exposure to externally provisioned microbes. Emergence from quiescence could also be associated with increases in colonization of microbes from the water column, leading to the observed increases in alpha diversity. Interestingly, in time point 7, when 50% of the colonies had emerged from quiescence, the corals that had already emerged from quiescence had higher alpha diversity than the corals still in quiescence, suggesting that quiescence directly influences the observed changes in diversity. Further experimental work is needed to understand how quiescence emergence (e.g., mucus production, colonization from the water column) versus initiation of feeding influences the increase in diversity on corals emerging from quiescence.

Unexpectedly, coral microbiomes exhibited consistent levels of dispersion throughout the quiescence period in both the active and present microbial communities. This was surprising, as tropical corals, and other animal hosts exposed to stressors, often show increased variability during a stressor event (the Anna Karenina hypothesis [26, 27]). The lowered dispersion before quiescence is in contrast to data that had been previously collected seasonally, in which spring-collected (e.g., after quiescence) corals show decreased variability in their microbiomes relative to those collected during other seasonal time points (15) . We suggest that this difference could be due to variation in the locations sampled (Rhode Island versus Massachusetts), the lower resolution of sampling timing in the previous study, idiosyncratic differences in the environment associated with the day of sampling for the Rhode Island corals, or higher B. psygmophilum densities in the corals in previous studies. Previous research suggests that B. psygmophilum can influence microbiome beta dispersion after disturbance (laboratory antibiotic treatment) (24), so perhaps the near absence of B. psygmophilum across all of the corals in this study explains the consistent beta dispersion levels among the samples. Here, our sampling times encompassed multiple time points, including those in which some corals were in quiescence, and some were not (T7), 1 week after corals had emerged (T8) and 2 months later (T9), revealing that postquiescence is not always associated with lowered dispersion, and consistent levels of beta diversity (intercolony variability) may be an evolved trait of these corals.

Quiescence was associated with a reshuffling of the coral microbial community that persisted after the corals emerged from quiescence. In fact, few taxa were consistently associated with corals over the course of the sampling time because of the marked compositional shift that began during quiescence. The taxonomic shifts we observed before and after quiescence are similar to those previously documented in fall- and spring-collected corals (15). This concurrence suggests there may be predictable or cyclical patterns in the microbiome composition associated with dormancy timing that could be important for coral holobiont health. Additionally, reshuffling or shifts in the microbiome associated with dormancy are common among host-microbes, including diapaused copepods (28), parasitoid wasps (8), mosquitos (29), squirrels (6), and bears (7).

Among the hypotheses about the roles the microbiome plays in dormancy are (i) compositional shifts that lead to the removal of pathogens and/or (ii) the replacement or maintenance of critical functions (30). Here, we see evidence for these two hypotheses based on the identity of the taxa and predicted functions.

Among the taxon changes that are associated with community composition shifts are the loss of copiotrophic, and particularly, pathogen-associated, bacteria as corals undergo quiescence. These taxa include Arcobacter, Pseudomonas, and taxa in the order Rickettsiales, including a taxon with 97.6% sequence identity to tropical coral parasite, “Ca. Aquarickettsia rohweri” (31). Indeed, many of these taxa are associated with diseases in tropical corals (32–34). We also generally observed a decrease in copiotrophs, including Endozoicomonas, a putative beneficial symbiont in many hosts (35). In tropical corals, Endozoicomonas decreases in response to thermal stress (36), suggesting that it may be released when a host is stressed (e.g., an adaptive response) (37). Because Endozoicomonas tends to have large genomes (4 to 6 Mb) (38), they are likely energetically costly to maintain in symbiosis (39), which may be why they are reduced during dormancy (and other stressor events).

We suggest that the loss of copiotrophic bacteria and putative pathogens associated with the beginning of dormancy is potentially a mechanism or a consequence of a period with limited resources, when the holobiont cannot support energetically costly microbes. Thus, this loss is the result of either these microbes voluntarily or passively leaving the coral’s microbiome or an active ejection by the coral. Alternatively, a decline in A. poculata holobiont metabolism may trigger a concomitant decrease in the production of molecules that enrich for specific bacterial associates. Interestingly, the lone dormancy-only-associated microbe was most closely related to UBA10353, a bacterial group that produces pederin, a bioactive polyketide, in sponges (40). A resulting hypothesis is that the increase in this bacterium could result in production of antimicrobial compounds to reduce the colonization of microbes while the host is quiescent and/or help explain the loss of microbes associated with dormancy. In contrast with previous conclusions from seasonal characterization of A. poculata microbiomes (15), the higher-resolution sampling in this study reveals that putative pathogens are not higher in proportional abundances in winter months, as previously described (41) but, rather, are at their maximum just before entry into quiescence and are then shed after initiation of quiescence.

During dormancy we observed an increase and maintenance of microbes associated with ammonia oxidation, nitrification, and nitrogen fixation, suggesting that the microbiome plays a role in the maintenance and acquisition of nitrogen while corals are not actively feeding. Among the taxa that likely contribute to replacing host functions were archaea in the genus “Ca. Nitrosopumilus,” a known associate of Astrangia poculata (15, 24), and bacteria in the order Nitrosococcales, (Cm1-21, MSB-1D1).

“Ca. Nitrosopumilus” and Nitrosococcales are common ammonia oxidizers (42). Here, corals are not actively feeding and do not have any visible algal symbionts; thus, these ammonia oxidizers may play an important role in nitrate acquisition for the host or other essential microbes. Corals in late quiescence and early nonquiescence states also show increases in nitrate-reducing Rhizobiales (Psuedahrensia and Filomicrobium) (43) and Magnetospira, a likely nitrogen fixer (44). These taxa, along with the ammonia oxidizers, suggest that the microbial community likely continues to bolster nitrogen cycling in the host as corals emerge from quiescence and may help build energetic reserves that were depleted during quiescence (14). The presence of these bacteria and archaea, particularly the ammonia oxidizers, during dormancy and after dormancy may help explain some of the acquisition of nitrogen (i.e., ammonia, the host’s preferred dissolved inorganic Nitrogen, DIN, source) for the host in general, which is usually attributed to heterotrophy and enhanced by algal symbionts (14, 45). For example, the increase in nitrifying microbes may explain the higher δ¹⁵N values in A. poculata tissues previously found in the winter compared to the fall (14), as winter corals are quiescent and do not rely on heterotrophy (or photoautrophy). The nutrients in the water column (lowered during the winter months) also suggest that external provisioning is less likely at this time, increasing the importance of the potential microbial contribution to nutrient cycling. More research is needed to understand the role of the microbiome in the cycling of nitrogen in coral tissues and how this may impact coral fitness after quiescence.

Corals in late quiescence and after emergence also showed increases in Flavobacteriales (e.g., Pseudofulvibacter, Ulvibacter; Fig. 4, Fig. S2 and S5). As microbial heterotrophs, it is possible that these taxa play a role in carbon cycling on and/or with the host before and just as the coral begins to feed actively. For example, in sponges with high heterotrophic microbial loads, evidence suggests that microbes play a role in dissolved organic matter (DOM) assimilation (46). Alternatively, the increase in Flavobacteriales may be due to increases in food availability and is not necessarily host-associated (47).

Replacement of host nutrition during dormancy is a common theme in host-microbial systems. In ground squirrels (Ictidomys tridecemlineatus), the gut microbiome plays a critical role during hibernation to recycle nitrogen (from urea), which supports tissue growth while the animal is not feeding (9), which is evolutionarily advantageous leading up to the breeding season. Furthermore, diapausing Daphnia eggs are enriched in Nitrospira bacteria, suggesting that nitrification may be a function of dormant microbiomes in other hosts (48). In parasitoid wasps, microbiomes are responsible for synthesizing glucose for nutrition during dormancy (8). In both mosquitos and bears, microbes are suggested to also play a role in host provisioning and lipid storage (7, 29). Nutritional provisioning, particularly of nitrogen, during dormancy is potentially a convergent trait across host and microbes that undergo dormant periods.

Predictable, seasonal quiescence in A. poculata has been observed and documented in the field (11), and in the lab, quiescence can be experimentally triggered by lowering temperatures to 5°C (Sean Grace, personal communication; 13). Emergence from dormancy was similarly found by raising temperatures above 5°C (13). Here, our findings support that the onset of dormancy was associated with temperatures reaching 5°C, suggesting winter temperature as one of the environmental triggers for the onset of dormancy. Indeed, corals experienced temperatures as low as 2°C (Fig. 1). Because temperature appears to play a role in dormancy and microbial dynamics, as oceans warm, there likely will be consequences for the timing and duration of dormancy in northern populations of A. poculata, the microbial shifts surrounding dormancy, and coral host physiology. However, absolute temperature is likely not the only trigger for emergence: corals began to emerge from dormancy while water temperatures remained close to 5°C (Fig. 1, Table 1); thus, there may be other factors that lead to the cessation of quiescence. An additional hypothesis is that nutrient availability may influence dormancy. Nutrients (e.g., ammonia, phosphate) are lower during the winter. As the microbiome shifts swiftly during quiescence, it is possible there is an interplay between the environment, the host, and the microbiome that triggers the onset of and emergence from quiescence. Further experimental work, including isolating the effects of temperature and season, and characterizing the mucus metabolome throughout seasonal and environmental shifts, will help to elucidate the relationship between microbial shifts and quiescence. Because A. poculata are facultatively symbiotic with microalgae and can also be used in aquarium-based studies (24), it is an ideal marine experimental system for investigation of the role of the microbiome in animal host dormancy.

Our findings suggest that A. poculata quiescence is involved in reshuffling the microbiome, leading to a new, persistent microbiome community structure. This shuffling is associated with shedding of potential pathogens, such as Rickettsiales, and a shift in the taxa that likely replace nutrition, such as “Ca. Nitrosopumilus,” while the host is inactive (Fig. 5). Overall, this study demonstrates that key microbial groups are related to quiescence in A. poculata and may play an indirect or direct role in the onset of and emergence from dormancy. Further understanding of the interactions between the coral and these specific microorganisms during this change in coral metabolic status will advance our understanding of coral host-microbiome dynamics and dormancy of host and microbes in general.

From late October 2020 to May 2021 we collected distinct white, or aposymbiotic, colonies of A. poculata at each of nine time points (n = 10 per time point; Table 1, Fig. 1). Corals were collected on SCUBA at 18 m from pilings on the Woods Hole Oceanographic Institution’s Iselin dock (41°31′25.1″N 70°40′19.3″W). We selected aposymbiotic colonies to reduce the potential effects of the algal symbionts in our understanding of dormancy-related microbial shifts. Selected colonies showed no visible coloration in any of the polyps (or near absence of algae; Fig. 1). Corals were collected using a hammer and chisel and were frozen in liquid nitrogen vapors immediately upon surfacing from the dive.

During eight of the time periods (starting on 3 December), we collected water samples in a 5-L Niskin bottle triggered at depth (18 m). We subsampled this water for macronutrients (25 mL, frozen to −20°C), which were analyzed as described previously (49). We also collected samples for total organic carbon and total nitrogen (40 mL acidified with concentrated phosphoric acid, 75 μL; n = 4 per time point, except time point 3, in which n = 3). For seawater microbial community analysis, we filtered 500 mL of the collected water through a 0.22-μm Sterivex filter (Millipore Sigma, Burlington, MA; n = 4 per time point) using peristaltic pressure and then flash froze the filter in liquid nitrogen vapors.

Seawater temperature data were from station BZBM3 in Woods, Hole, MA, measured 1.7 m from the mean lower low water line (50). Temperature data were averaged by day over the time period of sampling (October to May).

For both coral and water samples, DNA and RNA were extracted using the Quick DNA/RNA mini prep plus kit with ZR BashingBeads (0.1 and 0.5 mm; Zymo Research, Carlsbad, CA). For coral samples, one polyp (including mucus, tissue, and skeleton) of each coral colony was removed with a sterilized chisel and hammer. Coral fragments were added to the bead tubes and then suspended in 800 μL of DNA/RNA Shield and bead-beaten for 10 min at top speed on a vortexer. We added proteinase K (15 μL of 20 mg μL⁻¹) to further break down cells and isolate the DNA and then continued with the rest of the steps in the manufacturer’s protocol, including the DNase step for the RNA portion of the sample. Extracted RNA was frozen at −80°C, and DNA was frozen in at −20°C until further analysis.

For water samples, we opened the plastic case of the Sterivex filter using a sterilized steel cutting implement, removed the filter using a sterilized razor, cut the filter into strips over a sterile petri dish, and placed the filter into the bead tube using sterile tweezers. We then added 1,000 μL of DNA/RNA Shield to the sample and bead tube. Then, 30 μL of proteinase K (20 mg μL⁻¹) was added before we continued with the manufacturer’s protocol. Extraction blanks, which included reagents but no samples (n = 3 for the water protocol, n = 6 for the coral protocol), were carried out as well for both RNA and DNA.

We converted RNA to cDNA for sequencing the active microbiome using the New England Biolabs (Ipswich, MA) ProtoScript II first-strand cDNA synthesis kit. We followed the standard protocol and used 2 μL of coral RNA and 6 μL of water RNA as the template.

We prepared DNA and cDNA for 16S rRNA gene sequencing of the V4 region using barcoded 515FY (51) and 806RB (52) primers that target bacteria and archaea with standard barcodes (53). The PCRs (25 μL) were performed in duplicate per sample and prepared using the high-fidelity (HF) Phusion master mix with HF buffer (12.5 μL/sample), dimethyl sulfoxide (DMSO; 0.75 μL/sample) (New England Biolabs), molecular-grade water (7.25 μL/sample), the primers (1.25 μL of each), and 1 μL of template. The thermocycler conditions were an initial denaturation step of 95°C for 2 mins, and then 30 cycles of 95°C (20s), 55°C (15s), and 72°C (5 min), and a final elongation step of 72°C for 10 min.

Each PCR was run on 1.5% agarose, and the correct band (determined by the location of the positive control, ~400 bp) was excised. Excised bands were extracted and purified with the MinElute gel purification kit (Qiagen, Inc., Germantown, MD). Purified PCR products were quantified with a Qubit device, diluted to 1 ng μL⁻¹, and then pooled at 5 ng of purified product per sample. Each pool contained negative PCR controls with no visible bands and a mock community (Even, low community B; BEI Resources). The pools (3 total) were sequenced on an Illumina MiSeq instrument with 250-bp paired-end sequencing.

With the demultiplexed forward and reverse sequences, we used the DADA2 pipeline (54) in R (55) for quality control, merging sequences, and assigning amplicon sequence variants (ASVs). Forward and reverse reads were visually inspected for quality with DADA2 and ggplot2 and to determine the cutoff values (the average number of base pairs of which quality scores fell below 30) in the filter and trim step with the following parameters: filterAndTrim(fnFs, filtFs, fnRs, filtRs, truncLen = c(240, 150), maxN = 0, maxEE = c(2), rm.phix = TRUE, compress = TRUE, multithread = TRUE). Error rates were computed and used for sequence inference in DADA2. Sequences were then merged, and ASV tables were created. Because of the size of the data set, error rates and ASV tables were created per MiSeq run, the tables were then merged, and chimeras were checked and removed. Taxonomy was assigned using the SILVA v132 training set (56, 57), and retrieval of taxa from mock communities was checked.

The taxon table, ASV table, and metadata table were loaded into phyloseq (58), where chloroplasts and mitochondria were removed. Using the decontam package (59), we removed contaminant taxa using the prevalence of taxa (at 0.01) in the negative controls (including PCR negatives, extraction kit blanks, and water filter blanks for both DNA and cDNA).

Alpha diversity was calculated using Hill numbers D⁰, D¹, and D², which correspond to richness (rarefied), exponentiated Shannon diversity, and the inverse Simpson index, respectively (60). Higher D values indicate more even and speciose communities. We estimated diversity indices using phyloseq.

To compare the variability in coral communities over time, we computed Bray Curtis dissimilarities on the relative abundances of taxa within a sample. We then quantified the distance from each sample point to the group’s centroid (beta dispersion) using the betadisper function in the vegan package (61).

We tested for significant differences in alpha diversity and dispersion using linear models, comparing dormancy timing (before, during, after), sample time (time points 1 to 9), and active/present microbiome (cDNA/DNA) in R (54). Residuals were visually inspected to meet assumptions of heteroscedasticity and normality. Significance was assessed using ANOVA from the car package (62). When necessary, post hoc tests (Tukey’s HSD) were used to evaluate differences among levels in treatments.

To examine compositional changes within Astrangia poculata’s microbiome throughout the phases of dormancy (before, at the onset, during, and after), we used Bray Curtis dissimilarity matrices based on relative abundance. We then compared the microbial community composition using PERMANOVA in the vegan package (61) with sample times (1 to 9) and the timing around dormancy (before, during, after) and active/present microbiome as factors. To understand the differences in community composition across time points and types of sample, we used pairwise comparisons with the EcolUtils package (63). Coral and water microbial communities were visualized on an NMDS plot.

We assessed which microbial taxa changed in relative abundance with respect to timing of dormancy (before versus during, before versus after) using corncob (64). This method uses a beta-binomial model on the counts (number of reads) for each ASV to determine which taxa are significantly enriched from a reference level (in this case, before dormancy) in different treatments and compares them iteratively. We compared the active (DNA) and present (RNA) microbiomes from the coral and water separately.

To determine which microbial taxa compose the core of Astrangia poculata, and how these taxa changed over time, we determined which taxa were present at 80% prevalence across all samples within a time point (18) with the microbiome package (65).

We hypothesized potential functions of the microbiome using a functional inference tool based on taxonomy. Although there are drawbacks to tools that predict function from taxonomy (66, 67), here, we took a conservative approach and used broad categorizations of functions using the FAPROTAX database (68) and microeco package (69) in R, which were generated from published metabolic and ecological functions and suggested for environmental data (70). We extracted the functions associated with ASVs that were determined to be significantly enriched based on the corncob analysis to understand potential functional shifts across the dormancy time periods in the active and present microbiomes.

Sequences are available on the NCBI Sequence Read Archive (SRA) BioProject PRJNA860933 under the accession numbers: SAMN29871893-SAMN29872136.