Semantic Map for Structural Bioinformatics

Scientists and bioinformaticians alike are provided an incredibly rich variety of resources to retrieve, analyze, transform, integrate, and visualize scientific data. While this opulence should benefit scientific discovery by allowing the selection of the most suitable technology to implement digital scientific protocols, users experience difficulty identifying what the most suitable resources are, locating the resources, and integrating them into an efficient protocol executable.

The aim of the Semantic Maps project is to facilitate the identification and location of bioinformatics resources. The approach exploits a domain ontology to classify resources with respect to their scientific meaning. Each data source (e.g., database) is mapped to the concepts it corresponds to (e.g., GenBank is mapped to Biological Sequence) whereas each application or tool is mapped to the conceptual relationship it corresponds to (e.g., BLAST is mapped to Similar_to).

barre

<Acknowledgement>

The first version of the Semantic Maps was the BioNavigation system designed and developed with Louiqa Raschid and Maria Esther Vidal, our collaborators at the University of Maryland , and by Kaushal Parekh to complete his Professional Master Program in Computational Biosciences at Arizona State University .

The second version of the Semantic Maps was developed by Hervé Ménager, Institut Pasteur, Paris, in collaboration with Pierre Tufféry, Equipe de Bioinformatique Génomique et Moléculaire, INSERM U726, Université Denis Diderot ( Paris 7ème).
It benefits from BioNavigation and ZWTM developed by Emmanuel Pietriga at INRIA, France. It was used to support the Semantic Map for Structural Bioinformatics.
pimp1read more

We are now exploring how to develop a Semantic Map for BioMoby services thanks to a supplement grant by the National Science Foundation. Maliha Aziz is working on this extension to complete her Professional Master Program in Computational Biosciences at Arizona State University .

barre

Prototypes

Arrow down
BioNavigation Software Download

(Installation Requirements:
At least 25MB of disk space for the program folder)

 

arrow down
Download Installation Instructions for BioNavigation prototype interface
(.pdf)


      

Under Progress

Display of the semantic map - A semantic map is a complex multi-layered graph. First the domain ontology is a complex graph with directed labeled multi-edges and a isa hierarchy. Then each resource is mapped to the labeled edge it implements. The size and complexity of the graph requires the development of efficient graphical display tool, as well as new visualization strategies. We aim at developing a view mechanism to support the display of semantic maps.

Exploring the semantic map – The semantic map view mechanism will also support user interactions with the map to customize resource exploration.

Querying the semantic map – Resource discovery requires the identification of resources that implement specific scientific tasks, the location of the suitable resources. A scientific task may be implemented by a single resource or by connected resources. A scientific task may correspond to a resource path in the semantic map. A resource path may be of length 1 when a single resource implements the scientific tasks or of length n when n resources needs to be connected to implement the task. More complex problem may require a resource workflow (or sub-graph) for their implementation. We aim at developing a query language that expresses resource discovery queries. We are currently implementing a tool to navigate through the ontology. The overall size and complexity of the ontology poses several problems of display. We are using the ZWTM API to implement an improved interface. A (very alpha) prototype is available here.

Suitability of resources – In the Semantic Map approach a resource (resp. resource path or workflow) is suitable when its input and output corresponds to the input and output concepts of the scientific task. This captures a semantic suitability however other constraints exist when selecting a bioinformatics resource to implement a scientific task. Two resources may be mapped to the same conceptual relationship but produce different results. This problem is addressed in the Benchome project. Once a path or workflow of bioinformatics resources has been selected, the syntactical mapping of output/input in order to produce an executable workflow remains to be addressed. This is the aim of the BioOnMap project. Finally scientific protocols can be stored in ProtocolDB.

Service registration – We aim at automating the registration of the services registered in BioMoby and in Taverna. However, the information providing in the resource description does not able a straightforward registration in a semantic map. We invite providers of services devoted to structural bioinformatics to register their services in the semantic map.


Home - Project - News - Project Members - Sponsor - Papers - Contacts

Copyright © 2006-2008 SemanticMaps
All Rights Reserved