In this paper, the platform CALLISTO (CALmip Launches an Interface for a Semantic Toolbox Online, https://callisto.calmip.univ-toulouse.fr/) is presented, aiming at open science construction around semantic description of scientific data, software and papers. This work aims tackling the needs for a fine-grained description of multi-disciplinary scientific collaborations, from data creation to cross-fertilization. A framework for hosting comprehensive and semantic metadata for multi-disciplinary scientific teams is proposed. The platform provides a user-friendly online interface for data description, browsing and access. This research has been conducted in collaboration with the multi-disciplinary dataNooS academic alliance. CALLISTO is developed and maintained at the computing centre CALMIP.
The constant growth of scientific data available online and the growing supply of processing software can paradoxically make the reuse of data and algorithms harder. The lack of widelyused fine-grained description of scientific data, the need for explicit and ergonomic online interfaces for finding data are some of the challenges ahead towards open science. These concerns are at the heart of the work to encourage the widest possible adoption of the FAIR (Findable, Accessible, Interoperable and Reusable) principles. The simultaneous eforts at a national and European level to adopt the FAIR principles, as well as the existing dificulties within scientific projects in managing multidisciplinary data at several reading levels, contribute to raising scientists’ awareness of the issues involved in the management and use of their data. Within the University of Toulouse, an academic alliance has been created and named dataNooS1. It involves researchers and engineers from diferent scientific disciplines, whose main role is to study solutions for eficient data handling, interoperability and re-usability generalizable to future projects. As a member of dataNooS, the computing centre CALMIP developed and CEUR Workshop Proceedings maintains CALLISTO2 for ”CALmip Launches an Interface for a Semantic Toolbox Online”, implementing ideas from dataNooS3 in a production-level site. At the first step, the goal of CALLISTO consists of sharing data among partners with heterogeneous scientific backgrounds, ensuring their description and easing reuse in a secured calculation environment. At a second step, the objective is to open data, results and analysis processes to the public. CALLISTO is designed with a user-driven development process, through the study of data challenges coming from interdisciplinary projects such as the H2020 project Smart Morphing and Sensing for aeronautical configurations (SMS, http://smartwing.org/SMS/EU/) funded by European Commission (project No. 723402) and HiperBorea4, funded by the French National Research Agency (ANR) (project ANR-19-CE46-0003-01). These projects illustrate the challenges for scientific projects concerning the sharing and the description of heterogeneous and multidisciplinary data.
In this paper, CALLISTO is presented as an illustration of how the FAIRness of scientific data
can be achieved, and supplemented by other aspects such as descriptions of the software and
algorithms used to create, analyse or transform the data. This article presents creation of a
platform for linking data to software, scientific papers and arguments through the adoption of
well-known ontologies such as Micropublications [
SMS for ”Smart Morphing and Sensing” is the European project mentioned in the introduction. This project is a collaboration between experts with diferent scientific backgrounds (aeronautics, fluid mechanics, structural mechanics, control system...). Diferent teams from diferent countries worldwide make their own experiments in their competency areas, and then need to agree on common concepts in order to work with one another. The software libraries used by the various teams may be shared, and projects papers shall be understood by all the project’s stakeholders. A major requirement is then to have a common metadata set that all stakeholders understand, and all future users will also understand. Meanwhile, metadata shall also be able to express very narrow elements for specialists.
The first element to ease data sharing among project members is to provide a shared space
with storage space and metadata browsing. This is provided by Dataverse, which is used as a
metadata catalog and data registry in CALLISTO. An important element for experiments reuse
and reproducibility, is a good understanding of the analysis processes the data went through [
Specify the analysis process the data went through Link data to papers, claims and
arguments Access through user-friendly interfaces with no technical prerequisite
R A
Ontologies, following the widely-used Gruber definition are to be considered as ”explicit
speciifcation of a conceptualization”[
The goal pursued in CALLISTO by the means of software description is quite diferent from
software registries found in astrophysics 8, geosciences9 or domain-agnostic registries such as
Software Heritage [
By combining SWO in an ontology and data access description with domain knowledge, it is possible to go as deep as the languages used by the software, the libraries, and the diferent algorithms with their functionalities. This is very useful especially in science, where many algorithms in a single software may accomplish the same functionality (e.g., solving an equation) with diferent implementations. In CALLISTO, these formal descriptions rely on the ARCAS ontology that serves as a single register for these diferent elements. By reasoning on these description the system builds processing chains (workflows) adapted to the users’ requests when using the platform.
CALLISTO is the platform for collaborative science proposed by CALMIP. CALLISTO is available for public use15 and the underlying ARCAS ontology is downloadable16. The SMS example of using ARCAS in a scientific project is also available online 17. Public use of CALLISTO is restricted to items (data, software, ontologies...) made public by CALMIP users.
The primary intent of CALLISTO is to provide eficient data sharing, allowing for fine-grained
description of data and the scientific context they belong to, and that is realized by using
project-scale ontologies. All the afore mentioned ontologies are based on a common core,
the generic, domain-independant ontology called ARCAS (ARming CAllisto with Semantics).
Modeling the treatments associated with the data, whether for their creation or for future
analysis, is important. It provides information on the technical and scientific provenance of
the data, and allows future processing to be considered based on the precise description of
algorithms, functions and data formats acceptable to the software. ARCAS uses to this end SWO
[
Dataverse18 is used by CALLISTO as a data repository. AllegroGraph and Gruf [ 20] allows for the visualization, query and update of the associated ontologies. To ensure homogeneity and complementarity between the repository and the ontologies, CALLISTO performs a bidirectional exchange between the ontologies and Dataverse: selected ontology elements can be passed to the Dataverse metadata sets and reversely, the descriptions of the datasets automatically populate the associated ontology. Thus, the user can populate metadata fields coming from the ontology to Dataverse, and supplement this metadata with concepts from the ontology. Figure 1 shows the functional overview of CALLISTO.
Thanks to the GEOS module, CALLISTO aims at providing to its users advanced functionalities such as automatic or semi-automatic services composition. These functionalities rely on Allegrograph as the triple-store storing CALLISTO ontologies.
The role of metadata is to describe the strongest common block of description agreed upon by the majority of project participants. Therefore, the minimum common concepts can be understood by all regardless of the scientific field. In our approach, metadata refers to ontology items, allowing the fields to be filled in with a degree of flexibility while referring unambiguously to domain entities. This is an advantage over a ”classic” metadata field that can sometimes be iflled in diferently depending on the user and whose exact meaning can become ambiguous. The conclusion of the above, is that ontologies and metadata need to be related to each other. Nevertheless, being able to reliably match the content of metadata with the ontological representation of a knowledge domain is a delicate problem, dificult to generalize and often implemented in an ”ad hoc” way. In CALLISTO this is resolved by specifying the metadata themselves in the domain ontology so that they become a part of the domain knowledge representation. ARCAS provides a simple way of expressing that part of its content should be part of a metadata scheme. It uses three elements to this end: the annotation property ”metadataLevel” and the two relations ”hasMetadata” and ”hasMandatoryMetadata”. The property ”metadataLevel” indicates that the element it annotates is to be considered as an element of the metadata model for the ontology content. Its values may be 1 or 2, a value of 1 indicating that the element is an optional metadata field, and 2 indicating that the element is a mandatory metadata field. Figure 2 shows an example of metadata expression, for the HiperBorea instance of ARCAS. HiperBorea aims quantifying the impact of climate change on boreal permafrost, and drilling is used for experimental measurements. The HiperBorea scientific leaders chose to follow GTN-P metadata modeling of boreholes. As a consequence, elements in HiperBorea ontology have been named to reflect this schema. A template for drilling individuals is available in the dedicated HiperBorea ARCADIE version, providing properties and relations for relevant drilling description. These elements are part of the metadata associated to the boreholes. Appropriate relations and properties have been created, using the ARCAS ”MetaDataLevel” properties, the ”hasMetadata” and ”hasMandatoryMetadata” relationships, thus allowing automated Dataverse-compliant metadata generation
The description of the software, together with the description of the scientific claims related to the data and the papers, are part of the ”V.R.E.” (Virtual Research Environment) ofered by CALLISTO, exposed in Figure 3. This V.R.E serves two main purposes: • Services composition including data access and data analysis in several steps (workflows).
This service composition, based on the algorithm developed for astrophysics [
In this Figure, we see a claim in the center, lower part. This claim ”is qualified by” two elements, NSMB[21] and Power spectral density (PSD). It also ”supports” a scientific paper. ”is qualified by” and ”supports” are two relations from Micropublications ontology. These two elements come from the domain description for SMS project. The PSD is then an output of a data analysis service, in this example calculated by the periodogram method. PSD and the service are linked through the relation ”isCombinedToParam”, combining the PSD and an aggregate defining the PSD together with its format and unit. This service outputs PSD, and has an input that is ”NSMB-simulated near trailing edge activation” which one can find in some of the datasets in CALLISTO Dataverse instance (e.g. ”FDV 601” dataset in this example). This dataset ”has query software” (a relation in the ontology coming from GEOS module) some software that automates its access. Therefore, it is possible to go from a claim in a scientific paper to ontological elements and from these elements to data analysis pipelines.
Interoperability between data and analysis software is ensured by the description of the
elements in the ontology. Aggregates in ARCAS follow the specification of GEOS module of
ASON [
An extensive view of the benefits of using CALLISTO in a scientific project is given in a datapaper coming from SMS project[23]. This paper covers the functionalities: Data sharing, workflows description and overall enhancements of data sharing among project stakeholders. This illustrates some use cases benefiting from the CALLISTO’s features. As an example, the contributions of this approach concerning the services composition is the following: During the elaboration of the first processing workflow, it was necessary to define, in the ontology, relevant metadata for the definition of data access services and data analysis. This metadata was then transmitted from the ontology to the Dataverse data repository interfaces. When registering new data, the correct description of the processing and the provision of the right metadata at the time of deposit allowed the automatic inclusion, when appropriate, of the datasets in the workflow. The following Table 2 summarizes the requirements outlined in section 3 and the way CALLISTO meets them. An extensive discussion about the diferent platforms for scientific FAIR data sharing would go out the scope of this paper, nevertheless we indicate on Table 2 some elements of comparison with the European research infrastructure ELIXIR (www.elixir-europe.org). In the colum ”ELIXIR”, a ”A” indicates that ELIXIR adresses the same requirement (with diferent technical solutions) where a ”NA” indicates that, to the best of our knowledge ELIXIR does not.
1 2
In this paper, the platform CALLISTO has been presented, which aims to support Open Science with the added value of incorporating Dataverse as datasets repository, Allegro as a Triple Store and scientific papers description by the means of Micropublications. Furthermore, CALLISTO has been improved by the adoption of a multi-disciplinary point of view, through the production of project-specific ontologies (ARCADIE) all inheriting from the same ARCAS core ontology. CALLISTO is a semantic platform for improving FAIR and multi-disciplinary scientific data sharing that relies, when possible, on proven solutions (Dataverse, well-known ontologies, Allegro) and proposes to unify these elements thanks to the ARCAS and ARCADIE ontologies. The aim of this work is to enable a better understanding of shared data, by including metadata definition elements directly in the ontology and a mechanism for transforming these ontological elements into a format compatible with data repository software, namely Dataverse. Through the use in particular of SWO and MicroPublications, CALLISTO allows describing not only the data, but also the ecosystem in which they are inserted (processing software, scientific publications...). The second important feature of CALLISTO evolution is the development of an online workspace, supported by reasoning using ontologies at the center of CALLISTO. Virtual research environment functionalities like a semi-automatic workflow generation, are available online. A reuse of r o - c r a t e ids planned for the description of workflows created inside CALLISTO by the users to enhance their reusability. Finally, an extensive study of the possibilities ofered by SD will be conducted. digital preservation, Journal of biomedical semantics 5 (2014) 25. [16] B. Smith, Ashburner, C. M., Rosse, et al., The obo foundry: coordinated evolution of ontologies to support biomedical data integration 25 (2007) 1251–1255. doi:h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / n b t 1 3 4 6 . [17] B. Smith, A. Kumar, T. Bittner, Basic Formal Ontology for Bioinformatics, IFOMIS Reports, 2005. [18] J. Ison, M. Kalaš, I. Jonassen, D. Bolser, M. Uludag, H. McWilliam, J. Malone, R. Lopez, S. Pettifer, P. Rice, Edam: an ontology of bioinformatics operations, types of data and identifiers, topics and formats, Bioinformatics 29 (2013) 1325–1332. URL: https://doi.org/10.1093/bioinformatics/btt113. doi:1 0 . 1 0 9 3 / b i o i n f o r m a t i c s / b t t 1 1 3 . a r X i v : h t t p s : / / a c a d e m i c . o u p . c o m / b i o i n f o r m a t i c s / a r t i c l e p d f / 2 9 / 1 0 / 1 3 2 5 / 7 1 0 0 7 5 / b t t 1 1 3 . p d f . [19] T. Lebo, S. Sahoo, D. McGuinness, K. Belhajjame, J. Cheney, D. Corsar, D. Garijo, S. Soiland
Reyes, S. Zednik, J. Zhao, Prov-o: The prov ontology, W3C recommendation 30 (2013). [20] Franzinc,semanticwebtechnologies http://www.franz.com/agraph/allegrograph/, ???? [21] Y. Hoarau, D. Pena, J. B. Vos, D. Charbonier, A. Gehri, M. Braza, T. Deloze, E. Laurendeau, Recent developments of the navier stokes multi block (nsmb) cfd solver, in: 54th AIAA Aerospace Sciences Meeting, 2016, p. 2056. [22] D. Martin, M. Burstein, D. McDermott, S. McIlraith, M. Paolucci, K. Sycara, D. L. McGuinness, E. Sirin, N. Srinivasan, Bringing semantics to web services with owl-s, World Wide Web 10 (2007) 243–277. [23] J.-B. Tô, T. Louge, C. Rouaix, C. Jimenez Navarro, A. Marouf, M. Braza, ”aerodynamic performance in transonic regime around an a320 airfoil by means of electroactive morphing through vibration and slight deformation of the near-trailing edge region at high reynolds number”, datapaper preprint (????). doi:h t t p s : / / d a t a v e r s e . c a l l i s t o . c a l m i p . u n i v - t o u l o u s e . f r / d a t a s e t . x h t m l ? p e r s i s t e n t I d = d o i : 1 0 . 4 8 5 3 1 / J B R U . C A L M I P / 1 9 X G R Y .