Elucidating ecological complexity: Unsupervised learning determines global marine eco-provinces.
Ontology highlight
ABSTRACT:
SUBMITTER: Sonnewald M
PROVIDER: S-EPMC7259926 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Word embedding has benefited a broad spectrum of text analysis tasks by learning distributed word representations to encode word semantics. Word representations are typically learned by modeling local contexts of words, assuming that words sharing similar surrounding words are semantically close. We argue that local contexts can only partially define word semantics in the unsupervised word embedding learning. Global contexts, referring to the broader semantic units, such as the document or paragraph where the word appears, can capture different aspects of word semantics and complement local contexts. We propose two simple yet effective unsupervised word embedding models that jointly model both local and global contexts to learn word representations. We provide theoretical interpretations of the proposed models to demonstrate how local and global contexts are jointly modeled, assuming a generative relationship between words and contexts. We conduct a thorough evaluation on a wide range of benchmark datasets. Our quantitative analysis and case study show that despite their simplicity, our two proposed models achieve superior performance on word similarity and text classification tasks.
| S-EPMC7931948 | biostudies-literature
Project description:Phosphate (Pi) is an essential element to all living cells yet fluctuations in Pi concentrations are recurrent in the marine environment. Diatoms are amongst the most successful phytoplankton clade living in the world’s oceans adapting to and surviving long periods of sub-optimal conditions and resuming growth as soon as nutrient concentrations permit. The knowledge of the molecular underpinnings of diatom ecological success is, however, still very incomplete. Using high-throughput RNA sequencing we have identified the dynamics of the model diatom Phaeodactylum tricornutum global transcriptome in response to Pi fluctuations. We report genes encoding previously unidentified putative Pi transporters that, together with the alkaline phosphatases genes highly expressed under Pi depletion, are probably accountable for a very efficient Pi scavenging system. Our data also reveal the complexity of the P. tricornutum responsive sensory and signaling system that combines bacterial two-component systems with more complex pathways reminiscent of metazoans. This intricate set of membrane receptors and signal-transduction related proteins is likely responsible for the detection of environmental nutrient fluctuations, triggering a multitude of signaling cascades leading to gene regulatory events that define new adaptive physiological states. This together with the novel detection of a multitude of long non-coding intergenic RNAs specifically expressed under Pi stress, begins to provide insights into the complex molecular regulatory program involved in the resilience and ecological success of diatoms.
2016-01-01 | GSE66997 | GEO
Project description:People value the existence of a variety of marine species and habitats, many of which are negatively impacted by human activities. The Convention on Biological Diversity and other international and national policy agreements have set broad goals for reducing the rate of biodiversity loss. However, efforts to conserve biodiversity cannot be effective without comprehensive metrics both to assess progress towards meeting conservation goals and to account for measures that reduce pressures so that positive actions are encouraged. We developed an index based on a global assessment of the condition of marine biodiversity using publically available data to estimate the condition of species and habitats within 151 coastal countries. Our assessment also included data on social and ecological pressures on biodiversity as well as variables that indicate whether good governance is in place to reduce them. Thus, our index is a social as well as ecological measure of the current and likely future status of biodiversity. As part of our analyses, we set explicit reference points or targets that provide benchmarks for success and allow for comparative assessment of current conditions. Overall country-level scores ranged from 43 to 95 on a scale of 1 to 100, but countries that scored high for species did not necessarily score high for habitats. Although most current status scores were relatively high, likely future status scores for biodiversity were much lower in most countries due to negative trends for both species and habitats. We also found a strong positive relationship between the Human Development Index and resilience measures that could promote greater sustainability by reducing pressures. This relationship suggests that many developing countries lack effective governance, further jeopardizing their ability to maintain species and habitats in the future.
| S-EPMC3623975 | biostudies-literature
Project description:BackgroundEcological attributes estimated from food web models have the potential to be indicators of good environmental status given their capabilities to describe redundancy, food web changes, and sensitivity to fishing. They can be used as a baseline to show how they might be modified in the future with human impacts such as climate change, acidification, eutrophication, or overfishing.MethodologyIn this study ecological network analysis indicators of 105 marine food web models were tested for variation with traits such as ecosystem type, latitude, ocean basin, depth, size, time period, and exploitation state, whilst also considering structural properties of the models such as number of linkages, number of living functional groups or total number of functional groups as covariate factors.Principal findingsEight indicators were robust to model construction: relative ascendency; relative overhead; redundancy; total systems throughput (TST); primary production/TST; consumption/TST; export/TST; and total biomass of the community. Large-scale differences were seen in the ecosystems of the Atlantic and Pacific Oceans, with the Western Atlantic being more complex with an increased ability to mitigate impacts, while the Eastern Atlantic showed lower internal complexity. In addition, the Eastern Pacific was less organised than the Eastern Atlantic although both of these systems had increased primary production as eastern boundary current systems. Differences by ecosystem type highlighted coral reefs as having the largest energy flow and total biomass per unit of surface, while lagoons, estuaries, and bays had lower transfer efficiencies and higher recycling. These differences prevailed over time, although some traits changed with fishing intensity. Keystone groups were mainly higher trophic level species with mostly top-down effects, while structural/dominant groups were mainly lower trophic level groups (benthic primary producers such as seagrass and macroalgae, and invertebrates). Keystone groups were prevalent in estuarine or small/shallow systems, and in systems with reduced fishing pressure. Changes to the abundance of key functional groups might have significant implications for the functioning of ecosystems and should be avoided through management.Conclusion/significanceOur results provide additional understanding of patterns of structural and functional indicators in different ecosystems. Ecosystem traits such as type, size, depth, and location need to be accounted for when setting reference levels as these affect absolute values of ecological indicators. Therefore, establishing absolute reference values for ecosystem indicators may not be suitable to the ecosystem-based, precautionary approach. Reference levels for ecosystem indicators should be developed for individual ecosystems or ecosystems with the same typologies (similar location, ecosystem type, etc.) and not benchmarked against all other ecosystems.
| S-EPMC3998982 | biostudies-literature
Project description:Phosphate (Pi) is an essential element to all living cells yet fluctuations in Pi concentrations are recurrent in the marine environment. Diatoms are amongst the most successful phytoplankton clade living in the worldâs oceans adapting to and surviving long periods of sub-optimal conditions and resuming growth as soon as nutrient concentrations permit. The knowledge of the molecular underpinnings of diatom ecological success is, however, still very incomplete. Using high-throughput RNA sequencing we have identified the dynamics of the model diatom Phaeodactylum tricornutum global transcriptome in response to Pi fluctuations. We report genes encoding previously unidentified putative Pi transporters that, together with the alkaline phosphatases genes highly expressed under Pi depletion, are probably accountable for a very efficient Pi scavenging system. Our data also reveal the complexity of the P. tricornutum responsive sensory and signaling system that combines bacterial two-component systems with more complex pathways reminiscent of metazoans. This intricate set of membrane receptors and signal-transduction related proteins is likely responsible for the detection of environmental nutrient fluctuations, triggering a multitude of signaling cascades leading to gene regulatory events that define new adaptive physiological states. This together with the novel detection of a multitude of long non-coding intergenic RNAs specifically expressed under Pi stress, begins to provide insights into the complex molecular regulatory program involved in the resilience and ecological success of diatoms. Examination of Pi-responsive transcriptoms in diatoms.
2016-07-03 | E-GEOD-66997 | biostudies-arrayexpress
Project description:We consider the problem of performing unsupervised learning in the presence of outliers - that is, observations that do not come from the same distribution as the rest of the data. It is known that in this setting, standard approaches for unsupervised learning can yield unsatisfactory results. For instance, in the presence of severe outliers, K-means clustering will often assign each outlier to its own cluster, or alternatively may yield distorted clusters in order to accommodate the outliers. In this paper, we take a new approach to extending existing unsupervised learning techniques to accommodate outliers. Our approach is an extension of a recent proposal for outlier detection in the regression setting. We allow each observation to take on an "error" term, and we penalize the errors using a group lasso penalty in order to encourage most of the observations' errors to exactly equal zero. We show that this approach can be used in order to develop extensions of K-means clustering and principal components analysis that result in accurate outlier detection, as well as improved performance in the presence of outliers. These methods are illustrated in a simulation study and on two gene expression data sets, and connections with M-estimation are explored.
| S-EPMC3716393 | biostudies-literature
Project description:We address the problem, fundamental to linguistics, bioinformatics, and certain other disciplines, of using corpora of raw symbolic sequential data to infer underlying rules that govern their production. Given a corpus of strings (such as text, transcribed speech, chromosome or protein sequence data, sheet music, etc.), our unsupervised algorithm recursively distills from it hierarchically structured patterns. The adios (automatic distillation of structure) algorithm relies on a statistical method for pattern extraction and on structured generalization, two processes that have been implicated in language acquisition. It has been evaluated on artificial context-free grammars with thousands of rules, on natural languages as diverse as English and Chinese, and on protein data correlating sequence with function. This unsupervised algorithm is capable of learning complex syntax, generating grammatical novel sentences, and proving useful in other fields that call for structure discovery from raw data, such as bioinformatics.
| S-EPMC1187953 | biostudies-literature
Project description:The microbiome emits informative signals of biological organization and environmental pressure that aid ecosystem monitoring and prediction. Are the many signals reducible to a habitat-specific portfolio that characterizes ecosystem health? Does an optimally structured microbiome imply a resilient microbiome? To answer these questions, we applied our novel Eco-Evo Mandala to bacterioplankton data from four habitats within the Great Barrier Reef, to explore how patterns in community structure, function and genetics signal habitat-specific organization and departures from theoretical optimality. The Mandala revealed communities departing from optimality in habitat-specific ways, mostly along structural and functional traits related to bacterioplankton abundance and interaction distributions (reflected by ϵ and λ as power law and exponential distribution parameters), which are not linearly associated with each other. River and reef communities were similar in their relatively low abundance and interaction disorganization (low ϵ and λ) due to their protective structured habitats. On the contrary, lagoon and estuarine inshore reefs appeared the most disorganized due to the ocean temperature and biogeochemical stress. Phylogenetic distances (D) were minimally informative in characterizing bacterioplankton organization. However, dominant populations, such as Proteobacteria, Bacteroidetes, and Cyanobacteria, were largely responsible for community patterns, being generalists with a large functional gene repertoire (high D) that increases resilience. The relative balance of these populations was found to be habitat-specific and likely related to systemic environmental stress. The position on the Mandala along the three fundamental traits, as well as fluctuations in this ecological state, conveys information about the microbiome's health (and likely ecosystem health considering bacteria-based multitrophic dependencies) as divergence from the expected relative optimality. The Eco-Evo Mandala emphasizes how habitat and the microbiome's interaction network topology are first- and second-order factors for ecosystem health evaluation over taxonomic species richness. Unhealthy microbiome communities and unbalanced microbes are identified not by macroecological indicators but by mapping their impact on the collective proportion and distribution of interactions, which regulates the microbiome's ecosystem function.
| S-EPMC8625105 | biostudies-literature
Project description:Larger geographical areas contain more species-an observation raised to a law in ecology. Less explored is whether biodiversity changes are accompanied by a modification of interaction networks. We use data from 32 spatial interaction networks from different ecosystems to analyse how network structure changes with area. We find that basic community structure descriptors (number of species, links and links per species) increase with area following a power law. Yet, the distribution of links per species varies little with area, indicating that the fundamental organization of interactions within networks is conserved. Our null model analyses suggest that the spatial scaling of network structure is determined by factors beyond species richness and the number of links. We demonstrate that biodiversity-area relationships can be extended from species counts to higher levels of network complexity. Therefore, the consequences of anthropogenic habitat destruction may extend from species loss to wider simplification of natural communities.
| S-EPMC7614050 | biostudies-literature
Project description:BackgroundThe backbone of the eukaryotic tree of life contains taxa only found in molecular surveys, of which we still have a limited understanding. Such is the case of Picozoa, an enigmatic lineage of heterotrophic picoeukaryotes within the supergroup Archaeplastida, which has emerged as a significant component of marine microbial planktonic communities. To enhance our understanding of the diversity, distribution, and ecology of Picozoa, we conduct a comprehensive assessment at different levels, from assemblages to taxa, employing phylogenetic analysis, species distribution modeling, and ecological niche characterization.ResultsPicozoa was among the ten most abundant eukaryotic groups, found almost exclusively in marine environments. The phylum was represented by 179 Picozoa's OTU (pOTUs) placed in five phylogenetic clades. Picozoa community structure had a clear latitudinal pattern, with polar assemblages tending to cluster separately from non-polar ones. Based on the abundance and occupancy pattern, the pOTUs were classified into four categories: Low-abundant, Widespread, Polar, and Non-polar. We calculated the ecological niche of each of these categories. Notably, pOTUs sharing similar ecological niches were not closely related species, indicating a phylogenetic overdispersion in Picozoa communities. This could be attributed to competitive exclusion and the strong influence of the seasonal amplitude of variations in environmental factors, such as temperature, shaping physiological and ecological traits.ConclusionsOverall, this work advances our understanding of uncharted protists' evolutionary dynamics and ecological strategies. Our results highlight the importance of understanding the species-level ecology of marine heteroflagellates like Picozoa. The observed phylogenetic overdispersion challenges the concept of phylogenetic niche conservatism in protist communities, suggesting that closely related species do not necessarily share similar ecological niches. Video Abstract.
| S-EPMC11373171 | biostudies-literature