Sequencing Tropical Marine Biodiversity

Understanding the patterns and processes that shape the diversity of life has long been a central part of ecology. Ecological marine science, however, has been limited by the fact that visual surveys are often time- and labor-intensive and require specific taxonomic expertise. The majority of marine species are also very small, cryptic (dwelling in burrows or crevices), and even often lack species names, further complicating the process. The development of high-throughput DNA sequencing, coupled with the development and expansion of genetic databases, however, have allowed for the rapid genetic identification of entire communities of organisms simultaneously through metabarcoding. Here, I compare how these methods perform in complement with traditional survey methods and show how they can be used to efficiently and effectively characterize marine biodiversity. First, I tested whether a metabarcoding survey of environmental DNA (eDNA) from seawater using metazoan cytochrome c oxidase I (COI) primers could identify differences in community composition among five adjacent habitats at 19 sites across the tropical Caribbean Bay of Almirante in Panama. This was paired with simultaneous diver fish surveys in order to compare the two methods. Environmental DNA concentrated from seawater samples revealed a tremendous diversity of animals (8,586 operational taxonomic units, OTUs), including many small taxa that would be undetected in traditional in situ surveys. The eDNA survey detected 43 fish species not seen in the visual diver surveys, and captured species not previously reported in regional databases. Environmental DNA also revealed significant differences in fish and invertebrate community composition across adjacent habitats and areas of the bay, driven in part by taxa known to be habitat-specialists or tolerant to wave action, demonstrating the ability of broad eDNA surveys to identify biodiversity patterns in the ocean. Second, I sampled 0.5 m by 0.5 m plots of dead thin leaf lettuce coral (Agaricia tenuifolia) across the Bay of Almirante in order to capture and describe spatial patterns of reef community diversity. I compared metabarcoding using the COI mitochondrial gene and the 18S ribosomal gene and traditional morphological survey methods. The 18S metabarcoding performed better at higher taxonomic levels but underestimated species-level diversity compared to COI. Across the study, COI metabarcoding found over 13 times the number of animal species than macroorganism surveys. Coral reef communities showed structure based on their geographical location within the bay, with this being more pronounced in small organisms found using metabarcoding than in macroorganisms or the sessile encrusting community. It is estimated that, when surveyed with COI metabarcoding, reefs within the region contain at least 9,453 metazoan OTUs—likely even more with increased sampling given that the OTU accumulation curves of these data do not reach a plateau. Finally, to complement the survey of coral reefs and help complete our picture of the marine biodiversity of the Bay of Almirante, I metabarcoded animal communities under 2 mm in body size and sessile animals on mangrove roots and seagrass blades as well as sediment samples from around coral reefs, mangrove stands, and seagrass blades. Out of the three ecosystems, coral reefs and mangrove roots hosted the greatest amount of non-sediment biodiversity. Based on regional metazoan OTU pool estimates, mangrove roots and seagrass blades within the region are likely home to at least 8,429 species and 4,757 species, respectively. As with the coral reef data, the OTU accumulation curves of these data did not plateau and therefore likely represent a minimum. A large proportion of taxa found in mangrove and seagrass samples were also present in other ecosystems, but the majority of taxa found in reef samples were not detected in any other sample type, highlighting both the potential biodiversity toll of continued coral reef degradation and the potential role of habitat connectivity in influencing these ecosystems. Sediment samples, however, had very large estimates of regional metazoan OTU diversity of all sample types, with coral sediments at 6,151 OTUs, mangrove sediments at 10,120 OTUs, and seagrass sediments at 7,377 OTUs. Effective, cheap, and scalable biodiversity assessments will be crucial to scaling up ecology over the next century. Utilizing high-throughput DNA sequencing methods such as metabarcoding along with techniques such as environmental DNA sampling will allow ecologists to sample bigger areas, cheaply and accurately. As shown here, incorporating molecular methods into biodiversity science is not only more efficient, but also necessary to capture a more complete picture of biodiversity