Le projet BETSI

Base de donnée et thésaurus sur les traits des invertébrés du sol.

Quel était le projet BETSI ?

  • Un projet français avec un consortium permanent d’environ 15 chercheurs en écologie des sols
  • L’acronyme signifie « Biological and Ecological Traits for Soil Invertebrates »
  • Le projet a débuté en février 2011 et s’est terminé en mai 2014, mais la base de donnée continue d’être alimentée !
  • Le projet a été financé par le CESAB/FRB (Centre de Synthèse et d’Analyse sur la Biodiversité / Fondation pour la Recherche sur la Biodiversité)

La faune de sol est remarquablement diverse, elle peut représenter jusqu’à 80 % de la biodiversité animale. Tous ces organismes interagissent fortement avec leur environnement et réalisent des fonctions indispensables, jouant un rôle majeur dans le fonctionnement du compartiment ‘sol’ des écosystèmes. Parfois ces organismes opèrent comme « des ingénieurs d’écosystème, » et leur impact sur les sols est autant physique que chimique et biologique. Ils modifient le sol, structurent et contrôlent la biodisponibilité en eau et en éléments chimiques qui sont utiles ou toxiques pour les autres organismes. Ils établissent des relations mutualistes avec la microflore et sont aussi, telles les communautés de vers de terre, des indicateurs de qualité du sol. Les pressions croissantes des activités humaines sur le sol et des changements climatiques menacent la biodiversité de sol. Par exemple, nous observons une diminution de 20 à 90% de la biomasse de ver de terre trois ans après la conversion d’une prairie en culture.

Le projet de BETSI a synthétisé et organisé des données sur les traits fonctionnels des invertébrés de sol. Les traits biologiques et écologiques sont les caractéristiques des organismes leur permettant de faire face à l’environnement. Ces informations sont très utiles de diagnostiquer comment les contraintes ont un impact sur les individus, les populations et les communautés. Utiliser des traits comme les indicateurs la qualité écologique est une approche très nouvelle et prometteuse. Ce projet a développé une base de données d’une portée sans précédent sur cette facette sous-étudiée de la biodiversité. Il en résulte une base de données des traits fonctionnels et des pistes sur la compréhension de la réponse des organismes du sol aux facteurs de pression anthropique comme l’utilisation de sol et la pollution. Le travail d’un consortium interdisciplinaire a permis de prendre en compte un large panel de ces organismes.

http://betsi.cesab.org/

betsi

More details about the BETSI project organisation:

In the BETSI project, questions concerned the understanding of soil invertebrate biodiversity determinants. The objectives of this project were the definition of the functional traits for soil invertebrates, their synthesis and the identification of traits that are relevant to understand responses to some stress and disturbance exerted on the soil invertebrates.

We have proceeded in four steps: (i) state-of-the-art and definition of a common vocabulary (thesaurus), (ii) identification and selection of relevant information, (iii) creation of a database and (iv) test of invertebrate response to various constraints and disturbances (Figure 1).

betsi

Figure 1 – Chronological and logical links between the actions defined in BETSI project.

1/ Definition of a common vocabulary (Pey et al. accepted)

As highlighted previously, the current use of traits leads to misunderstanding data integration and interpretation. Trait semantics are especially concerned. The integration of the trait data on soil invertebrates is a key issue which must be orientated towards semantic data integration. It deals with the heterogeneity of the terms employed (terminology) across time. It preserves their meanings (scientific concepts) and also captures their interrelationships. The thesaurus for soil invertebrate trait-based approaches (T-SITA) was constructed through a web-based tool, designed for the collaborative construction of thesauri in ecology: the Thesauform (Laporte et al. 2012).  T-SITA is a first initiative to deal with the semantics of traits and ecological preferences for soil invertebrates. The Thesauform allows the building of a thesaurus which resulted in a hierarchy of terms organized according to their meaning, i.e. their concept. Each term of the hierarchy is described by a defined number of its properties: preferred label, definition, bibliographic reference of the definition, abbreviation, synonym, related term and preferred unit. The building procedure is performed collaboratively by a scientific expert community. It is divided into three successive steps: editing, validation and supervision. The editing step consists in the opportunity for each scientific expert to (i) modify and enrich the properties of a term, (ii) modify the hierarchy, (iii) add or delete a term or (iv) add a comment. The validation phase consists of a voting procedure within the scientific expert community on the different amendments produced during the editing phase. At each of these first two steps, several scientific experts can access the Thesauform simultaneously. The supervision phase aims at controlling the semantic consistency of the votes. It is mainly done by the editors of the thesaurus before the release of the final version. The whole procedure described above can be repeated indefinitely to continually improve the semantic content of the thesaurus.

Identification and selection of relevant information
Before starting T-SITA editing phase, twenty-one experts in soil invertebrate ecology provided a list of approximately 80 well-known terms of traits and ecological preferences based on literature survey. These trait/ecological preference terms were first chosen because they are commonly used for at least four notable invertebrate taxonomic groups with different biological strategies: earthworms, ground beetles, spiders and springtails. Nevertheless, the thesaurus design is not limited to such soil invertebrate groups. It is possible to input trait terms for all soil invertebrates and/or specific trait terms for a single soil invertebrate group (e.g. collembolan ocelli number). Finally, T-SITA reflects the agreement of a scientific expert community to fix semantic properties (e.g. definition) of approximately 100 traits and ecological preferences.
Before the onset of BETSI project, a number of datasets had been listed. It must contained trait-species data or species-environment data. Some of them were already organized into RDBMS (Ponge and Salmon 2013), other in spreadsheets database and a majority of them were grey literature or handbooks.

2/ BETSI database

The current use of traits for soil invertebrates resulted from isolated initiatives. They produced quantities of disjointed heterogeneous data. Without efforts to integrate them, the emergence of new knowledge from combining, reusing or sharing them will remain scarce and time-consuming. Our main concern is to provide soil invertebrate scientists with tools which allow data identification, availability and interoperability. The semantic web offers such kinds of tools by being based on the key principles of metadata, controlled vocabularies and ontologies.
A database for collecting all kind of trait data has been built under the database management system PostgreSQL (http://www.postgresql.org/docs/9.1/static/reference.html). Database structure contains several modules to deal with the heterogeneous format of trait data:

  • a taxonomical module which deals with the soil fauna taxonomy.
  • plot and experimental modules which allow inputting accurate experimental data (soil, fauna, sample conditions, measured traits)
  • a trait module which allows inputting inaccurate literature data (textual and numerical trait data without any accurate described sample conditions)
  • finally a source module which allows assigning a source for every object of the database whatever the module it belongs.

Inside modules, data were described using common technical and scientific standards (e.g. NUTS code for town identification, Fauna Europeae for taxonomy). It offers the possibility to the database to be interoperable with any other ecoinformatic tool.  T-SITA has been successfully linked with it. Such a link enhances data management and improves data scientific integrity. In addition, intern structure of the modules present metacolumns allowing the scientific standards to be modified among time. This last property is crucial for the database content evolution.

Liste des participants (2011):

Participant Affiliation Ville Mot-clé 1 Mot-clé 2 Mot-clé 3
M Hedde (PI) INRA, UR 251 Versailles macrofauna database data analysis
B Pey (PstDoc) ENSAT, EcoLab Toulouse macrofauna mesofauna ecoinformatics
A Pasquet CNRS, UR AFPA Nancy spider behaviour predation
Y Capowiez INRA, UR1115 Avignon macrofauna beneficial agriculture
C Pelosi INRA, UR 251 Versailles earthworm agriculture functions
C Pernin Univ, EA4515 Lille collembola pollution agriculture
F Grumiaux Univ, EA4515 Lille earthworms pollution agriculture
J Nahmani CNRS, CEFE Montpellier macrofauna database mediterranean
S Salmon MNHN, UMR7179 Brunoy collembola database agriculture
F Dubs IRD, UMR IESS Bondy macrofauna mesofauna landscape
J Cortet Univ, CEFE Montpellier collembola pollution agriculture
T Decaëns Univ, CEFE Montpellier macrofauna barcoding biogeography
D Cluzeau CNRS, ECOBIO Paimpont earthworm agriculture database
A Auclerc Univ, UMR 1120 Nancy collembola reclamation macrofauna
M Guernion CNRS, ECOBIO Paimpont earthworm agriculture database
J-F Ponge MNHN, UMR7179 Brunoy collembola pollution agriculture