Data(base) integration

Data(base) integration (cont.)

• Databases are very different in structure, in content
• Lots of applications require managing several databases
  • after company mergers
  • combination of administrative data for e-Government
  • biochemical, genetic, pharmaceutical research
  • etc.
• Most of the public data are accessible from the Web; proprietary data may not be yet
  • this should be done for easier collaboration!

What Is Needed?

• (Some) data should be available for machines for further processing
• Data should be possibly combined, merged on a Web scale (also referred to as “data integration”)
• Machines may also need to reason about that data
• What we need is a “Web of Data”
• Let us walk through a simple example…

A simplifed bookstore data (dataset “A”)

1^st step: export your data as a set of relations

Some notes on the exporting the data

• Relations form a graph
  • the nodes refer to the “real” data or contain some literal
  • how the graph is represented in machine is immaterial for now
• Data export does not necessarily mean physical conversion of the data
  • relations can be generated on-the-fly at query time
    • via SQL “bridges”
    • scraping HTML pages
    • extracting data from Excel sheets
    • etc.
• One can export part of the data

Another bookstore data (dataset “F”)

2^nd step: export your second set of data

3^rd step: start merging your data

3^rd step: start merging your data (cont.)

3^rd step: merge identical resources

Start making queries…

• User of data “F” can now ask queries like:
  • «donnes-moi le titre de l’original» (ie: “give me the title of the original”)
• This information is not in the dataset “F”…
• …but can be automatically retrieved by merging with dataset “A”!

However, more can be achieved…

• We “feel” that a:author and f:auteur should be the same
• But an automatic merge doest not know that!
• Let us add some extra information to the merged data:
  • a:author same as f:auteur
  • both identify a “Person”:
    • a term that a community may have already defined:
      • a “Person” is uniquely identified by his/her name and, say, homepage
      • it can be used as a “category” for certain type of resources

3^rd step revisited: use the extra knowledge

Start making richer queries!

• User of dataset “F” can now query:
  • «donnes-moi la page d’accueil de l’auteur de l’original»
  • (ie, “give me the home page of the original’s author”)
• The data is not in dataset “F”…
• …but was made available by:
  • merging datasets “A” and datasets “F”
  • adding three simple extra statements as an extra “glue”
  • using existing terminologies as part of the “glue”

Combine with different datasets

• Using, e.g., the “Person”, the dataset can be combined with other sources
• For example, data in Wikipedia can be extracted
  • there is an active development to add some simple semantic “tag” to wikipedia entries
  • we tacitly presuppose their existence in our example…

Merge with Wikipedia data

Is that surprising?

• Maybe but, in fact, no…
• What happened via automatic means is done all the time, every day by the users of the Web!
• The difference: a bit of extra rigor (e.g., naming the relationships) is necessary so that machines could do this, too

It could become even more powerful

• We could add extra knowledge to the merged datasets
  • e.g., a full classification of various type of library data
  • geographical information
  • etc.
• This is where ontologies, extra rules, etc, may come in
• Even more powerful queries can be asked as a result

What did we do?

The abstraction pays off because…

• … the graph representation is independent on the exact structures in, say, a relational database
• … a change in local database schemas, XHTML structures, etc, do not affect the whole, only the “export” step
  • “schema independence”
• … new data, new connections can be added seamlessly, regardless of the structure of other data sources

So where is the Semantic Web?

• The Semantic Web provides technologies to make such integration possible!
• It is a suite of technologies for the abstract data model, querying the data, defining ontologies, taxonomies, etc.
  • I do not have time for the exact details here…

The “bio” domain and the Semantic Web: ontologies

• A number of ontologies have been developed already:
• the US Cancer Institute’s Cancer Ontology, the Gene Ontology, the BioPax Molecular Pathway Ontology, the SWAN Project for the Alzheimer Disease research community, bio-zen ontology in neuroscientific and biomedical research, BrainPharm (Pathological Mechanisms in Alzheimer's Disease) from Yale Univ., …
• These are available in the ontology language defined by W3C
  • huge and powerful “glues” in our example above!

The “bio” domain and the Semantic Web: data sets

• A number of data sets are being exposed:
• BrainPharm and SWAN data cited above, NIST’s data on Thermodynamics of Enzyme-Catalyzed Reactions, RDF version of UniProt, …
• Work is going on to develop general methods for further data exposures; see, eg:
  • http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/Data

Health Care and Life Sciences Interest Group

• There has a been a great interest in these technologies from a number of R&D groups
• W3C formed the “Health Care and Life Sciences Interest Group”
  • goal is to explore the feasibility of these technologies
  • build demonstrations, explore possibilities
  • group runs until end of 2007, next steps are being explored
• Participants include Merck, AstraZeneca, Pfizer,Teranode, Partners HealthCare, IBM, Oracle, Agfa, HP, Universities of Amsterdam, Manchester, Yale, …

Some work areas of the HCLS IG

• Various “task forces”:
  • Develop techniques to convert, export, access, etc, biomedical data for data integration
    • e.g., conversion of the Entrez Gene data into RDF (33GB of data, from XML to RDF; the size was reduced during conversion!)
  • Evaluate, facilitate, etc, the creation of core vocabularies and ontologies in the area, possibly develop usage patterns
  • Look at clinical pathways, accommodate them with both patient heterogeneity and evolving clinical context
  • “Drug Safety and Efficacy”, ie, integrating the various steps needed in, eg, in an FDA approval process, clinical trial planning, reporting, management, etc.
• Use cases/demonstrations are being developed on data integration; a Workshop is organized in Banff, Canada, in May 2007

Other examples

• “Online community for knowledge sharing between clinicians in oral medicine in Sweden” (by Marie Gustafsson and others)
• Application for Traditional Chinese Medicine (by Huajun Chen and others)
  • integration of over 70 databases with a shared ontology

Semantic Web Applications

• Data integration is but one paradigm of Semantic Web usage
• Some others include knowledge management, labelling (multimedia) data, content adaptation, semantically oriented search,…
• Lots of tools are at disposal; e.g., Oracle’s 10g database is prepared for Semantic Web data storage and integration

Conclusions

• The Semantic Web is there to integrate data on the Web
• The goal is the creation of a Web of Data
• A major new avenue for Health Care and Life Sciences

Thank you for your attention!

These slides are publicly available on:

http://www.w3.org/2007/Talks/0314-London-IH/

in XHTML and PDF formats; the XHTML version has active links that you can follow