Health Care and Life Science Rule Responder (HCLS)
Note that the current version of this use case is mainly intended to show the technical rule-based service integration and composition.
Hence we show the queries and answers intended for machine-consumption in Reaction RuleML format as platform-independent interchange format and (currently) do not provide a computational-independent graphical or natural language user interface. But this may change in future.
Description:
The Health Care and Life Science Rule Responder use case develops a rule-based eScience inference infrastructure in the domain of
Health Care and Life Science, which will dynamically integrate, by the use of rules and common query languages, existing Web-based scientifiy services and datasources such as
UniProt, GoPubMed, EMBL-EBI Patent Abstracts,
W3C HCLS Knowledge Base, SWAN, and many other services.
The aim is to provide a HCLS Rule Responder system which allows asking scientific queries to this emerging eScience infrastructure. Rules are used to describe
complex (expert) decision logic, reactive (behavioural) process logic and transformation logic.
Distributed rule-based inference services (Prova rule engines with local specialized rule bases) can be deployed
on a efficient and scalable service middleware providing enhanced rule-based access to the existing services and data sources on the web using more than 30 selectable transport protocols such as HTTP, JMS, SOAP etc. and common rule interchange format de-facto standard (Reaction RuleML)
The data is collected dynamically at run-time via the generic query built-ins of the used rule engine (Prova) which supports queries using common query languages such as SPARQL, RDF Triple, XPath/XQuery, SQL
and the native acces to Java object functions of Prova to use existing programatic procedural Java API functions in the rule code.
Following the idea of Web Services and Grid infrastructures this use case adresses the problem of rule-based heterogenous data integration, service composition and rule interchange by providing:
- a common platform-independent rule interchange language (Reaction RuleML http://ibis.in.tum.de/research/ReactionRuleML/) for the interchange of queries, answers and rules
- an efficient and scalable middleware (the Rule Responder framework) for deploying arbitrary rule engines (inference services) and web-based services (as in SOAs and grid infrastructures) and interchanging messages as asynchronous or synchronous events (complex event processing)
- platform-specific rule engines and inference services to intelligently answer queries and access the existing eScience services
In particular, in this use case we make use of Prova (http://www.prova.ws/), a highly expressive distributed Semantic Web rule engine which supports complex reaction rule-based workflows, rule-based complex event processing, distributed inference services, rule interchange, rule-based
decision logic and dynamic access to external data sources, web-based services and Java APIs. Prova follows the spirit and design of the recent W3C Semantic Web initiative and combines declarative rules, ontologies and inference with dynamic object-oriented programming and access to external data sources. One of the key advantages of Prova is its elegant separation of logic, data access, and computation and its tight integration of Java, Semantic Web technologies and enterprise service-oriented computing and complex event processing technologies.
Read more about the Health care and Life Science Use Case
Query Examples:
You might copy and paste the examples in the form above. The example queries are written in Reaction RuleML format and the results are also wrapped in this format (similar to XML SOAP for web services).
Research Questions |
Actions |
Search Results |
Knowledge discovered |
Query |
Which proteins may cause Alzheimer Disease (AD)? |
Search Uniprot for Alzheimer |
Returns several highly toxic proteins such as Amyloid beta A4 protein precursor (Alzheimer disease amyloid protein) that disrupt brain mechanisms for learning and memory |
ADDLs are a surprising new form of the amyloid beta protein |
example |
What is a good therapeutic target for AD? |
Search SWAN for target OR therap*(citesConcept TherapeuticTarget - see query 1 below) |
Returns 2 hypotheses [add specific statements] |
Beta amyloid in various forms (ADDL, oligomer, etc.) are targets |
example |
Which publications are about ADDLs? |
Search goPubMed for publications about ADDLs |
Returns (currently) 29 articles |
Possible therapeutic approaches aimed at lowering ADDLs in AD patients |
example |
What is the top location for Alzheimer disease reasearch addressing ADDLs as therapeutic target? |
Use the GoPubMed Statistics service to compute the statics based on the found publications |
Returns Evanston as the top location |
Most publications about ADDLs are coming from Evanston |
example |
Who is the top author in ADDLs research? |
Use the GoPubMed Statistics service to compute the statics based on the found publications |
Returns William Klein as top author |
Hypothesis: William Klein works in Evanston. Simple prove (calls Google service API): http://www.biochem.northwestern.edu/ibis/faculty/klein.htm |
example |
Which patents exists for ADDLs? |
Search EMBI-EBL patent abstract database for patents addressing ADDLs |
Return (currently) two patents from Klein William |
AMYLOID BETA-DERIVED DIFFUSIBLE LIGANDS (ADDLS), ADDL-SURROGATES, ADDL-BINDING MOLECULES, AND USES THEREOF |
example |
Who is the expert in ADDLs research? |
Use the expert finder rule to find the expert which combines the previous five queries 1-5 |
Return (currently) Klein William as expert since he has most publications and two patents in ADDLs research |
William Klein is the (current) expert in ADDLs reasearch as therapeutic target for AD |
example |
Other general queries for the HCLS Rule Responder are: (Signatures | Mode Declarations | Example)
- Signature describes the signature of the public rule functions, which might be queried.
- Mode declarations, describe the modes of the function's arguments: input (+), output (-), undefined (?).
- Example: an example (you might copy and paste the examples in the Rule Responder form):
Which interfaces (public rule functions) are provided by the rule responder inference service?
- interface(Signature, ModeDeclarations, Description) | interface(?, ?, ?) [example]
Which pragmatic performatives does this service understand?
- performative(Performative) | performative(?) [example]
Which specialized rule services are known and can be accessed by the HCLS rule responder?
- service(ServiceName,ServiceURL,ServiceOutputFormat) | service("?","?","?") [example]
Ask a query to a specialised service, e.g GoPubMed rule service, returning the data in the selected output format (if supported)? Note, the rule responder automatically delegates the query.
- retrieveData(Result, Service, Format, Query) | retrieveDate(-, +,+,+) [example]
More Example Queries
Sources:
HCLS Rule Responder Service:
GoPubMed Rule Responder Service:
W3C HCLS KB Rule Responder Service:
EMBL-EBI Patent Abstract Rule Responder Service:
Prova SPARQL Examples:
- Demonstrate built-in SPARQL integration in Prova (Prova Source)
|
Use Cases 
