Dataset Information

Integrating species and interactions into similarity metrics: a graph theory-based approach to understanding community similarity.

ABSTRACT: Community similarity is often assessed through similarities in species occurrences and abundances (i.e., compositional similarity) or through the distribution of species interactions (i.e., interaction similarity). Unfortunately, the joint empirical evaluation of both is still a challenge. Here, we analyze community similarity in ecological systems in order to evaluate the extent to which indices based exclusively on species composition differ from those that incorporate species interactions. Borrowing tools from graph theory, we compared the classic Jaccard index with the graph edit distance (GED), a metric that allowed us to combine species composition and interactions. We found that similarity measures computed using only taxonomic composition could differ strongly from those that include composition and interactions. We conclude that new indices that incorporate community features beyond composition will be more robust for assessing similitude between natural systems than those purely based on species occurrences. Our results have therefore important conceptual and practical consequences for the analysis of ecological communities.

SUBMITTER: Lopez DN

PROVIDER: S-EPMC6546078 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Integrating species and interactions into similarity metrics: a graph theory-based approach to understanding community similarity.

López Daniela N DN Camus Patricio A PA Valdivia Nelson N Estay Sergio A SA

PeerJ 20190531

Community similarity is often assessed through similarities in species occurrences and abundances (i.e., compositional similarity) or through the distribution of species interactions (i.e., interaction similarity). Unfortunately, the joint empirical evaluation of both is still a challenge. Here, we analyze community similarity in ecological systems in order to evaluate the extent to which indices based exclusively on species composition differ from those that incorporate species interactions. Bo ...[more]

PMID: 31183257

Similar Datasets

Project description:Drug-drug interactions are preventable causes of medical injuries and often result in doctor and emergency room visits. Computational techniques can be used to predict potential drug-drug interactions. We approach the drug-drug interaction prediction problem as a link prediction problem and present two novel methods for drug-drug interaction prediction based on artificial neural networks and factor propagation over graph nodes: adjacency matrix factorization (AMF) and adjacency matrix factorization with propagation (AMFP). We conduct a retrospective analysis by training our models on a previous release of the DrugBank database with 1,141 drugs and 45,296 drug-drug interactions and evaluate the results on a later version of DrugBank with 1,440 drugs and 248,146 drug-drug interactions. Additionally, we perform a holdout analysis using DrugBank. We report an area under the receiver operating characteristic curve score of 0.807 and 0.990 for the retrospective and holdout analyses respectively. Finally, we create an ensemble-based classifier using AMF, AMFP, and existing link prediction methods and obtain an area under the receiver operating characteristic curve of 0.814 and 0.991 for the retrospective and the holdout analyses. We demonstrate that AMF and AMFP provide state of the art results compared to existing methods and that the ensemble-based classifier improves the performance by combining various predictors. Additionally, we compare our methods with multi-source data-based predictors using cross-validation. In the multi-source data comparison, our methods outperform various ensembles created using 29 different predictors based on several data sources. These results suggest that AMF, AMFP, and the proposed ensemble-based classifier can provide important information during drug development and regarding drug prescription given only partial or noisy data. Additionally, the results indicate that the interaction network (known DDIs) is the most useful data source for identifying potential DDIs and that our methods take advantage of it better than the other methods investigated. The methods we present can also be used to solve other link prediction problems. Drug embeddings (compressed representations) created when training our models using the interaction network have been made public.

Project description:Bipolar disorder is often misdiagnosed as major depressive disorder, which leads to inadequate treatment. Depressed individuals versus healthy control subjects, show increased expectation of negative outcomes. Due to increased impulsivity and risk for mania, however, depressed individuals with bipolar disorder may differ from those with major depressive disorder in neural mechanisms underlying anticipation processes. Graph theory methods for neuroimaging data analysis allow the identification of connectivity between multiple brain regions without prior model specification, and may help to identify neurobiological markers differentiating these disorders, thereby facilitating development of better therapeutic interventions. This study aimed to compare brain connectivity among regions involved in win/loss anticipation in depressed individuals with bipolar disorder (BDD) versus depressed individuals with major depressive disorder (MDD) versus healthy control subjects using graph theory methods. The study was conducted at the University of Pittsburgh Medical Center and included 31 BDD, 39 MDD, and 36 healthy control subjects. Participants were scanned while performing a number guessing reward task that included the periods of win and loss anticipation. We first identified the anticipatory network across all 106 participants by contrasting brain activation during all anticipation periods (win anticipation + loss anticipation) versus baseline, and win anticipation versus loss anticipation. Brain connectivity within the identified network was determined using the Independent Multiple sample Greedy Equivalence Search (IMaGES) and Linear non-Gaussian Orientation, Fixed Structure (LOFS) algorithms. Density of connections (the number of connections in the network), path length, and the global connectivity direction ('top-down' versus 'bottom-up') were compared across groups (BDD/MDD/healthy control subjects) and conditions (win/loss anticipation). These analyses showed that loss anticipation was characterized by denser top-down fronto-striatal and fronto-parietal connectivity in healthy control subjects, by bottom-up striatal-frontal connectivity in MDD, and by sparse connectivity lacking fronto-striatal connections in BDD. Win anticipation was characterized by dense connectivity of medial frontal with striatal and lateral frontal cortical regions in BDD, by sparser bottom-up striatum-medial frontal cortex connectivity in MDD, and by sparse connectivity in healthy control subjects. In summary, this is the first study to demonstrate that BDD and MDD with comparable levels of current depression differed from each other and healthy control subjects in density of connections, connectivity path length, and connectivity direction as a function of win or loss anticipation. These findings suggest that different neurobiological mechanisms may underlie aberrant anticipation processes in BDD and MDD, and that distinct therapeutic strategies may be required for these individuals to improve coping strategies during expectation of positive and negative outcomes.

Dataset Information

Integrating species and interactions into similarity metrics: a graph theory-based approach to understanding community similarity.

Publications

Integrating species and interactions into similarity metrics: a graph theory-based approach to understanding community similarity.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure