Drug Discovery: An Innovative Approach to Application ... · compelled drug discovery to embrace a...

white paper

Drug Discovery: An Innovative Approach toApplication Integration using SOA & Ontology

Cognizant White Paper

Executive Summary

Service Oriented Architecture (SOA) has radicallychanged the application integration landscape.However, the need for intelligent servicediscovery based on invocation context andservice capability is still a challenge. Ontology-based techniques are being experimented with as

a possible answer for creatingsemantically enriched SOAs. In thispaper, we discuss the ongoingexperimentation with usingontology as a layer above Webservices to achieve semanticintegration. We have chosen anarea of drug discovery calledsample management where life

sciences companies face application integrationchallenges. We have created a conceptualontology model for life sciences for use in ourstudies. This paper first details the experimentprocess and then demonstrates how to achievesemantic integration using ontological means.

Introduction

The fundamental driver for any integrationeffort is the need to share information within theenterprise and with business partners.Increasingly it is becoming evident thatenterprises that have most effective informationsharing are the ones that remain competitiveand successful. Due to the large heterogeneity of

the IT landscape in most large enterprises,enabling application integration is not an easytask. Application software integration, as a field,has received considerable attention over thepast two decades as a result of expandingbusiness requirements (a product of functionalsilos built around disparate and disconnectedsystems) and technological advances aimed ateliminating this challenge, once and for all.

Businesses need to operate in a globalenvironment with the ability to rapidly respond tomarket situations. The ecosystem of partners,suppliers, vendors need to be seamlesslyintegrated in a highly flexible manner. ExtensibleMarkup Language (XML) has emerged as apowerful self-describing language to enablebusinesses to share information and conducttransactions on the Internet. The emergence ofXML as a standard, to a large extent, has driven theevolution of application integration technologies.

Primitive efforts in application integration werefocused on created specific point-to-pointsolutions that were usually custom built. Then,innovative middleware products came intobeing. These middleware products typicallyformed a central hub through which allintegration happened. Most recent advances tapService Oriented Architectures (SOA) to enableapplication integration7. SOA consists ofservices that offer interoperability capabilitiesbuilt into the network.

Ontology-based techniques are being

experimented with as apossible answer for

creating semanticallyenriched SOAs.

2 white paper

SOA can be considered as a business-centricapproach for enabling integration. Visibility,interaction and effect are the key concepts inany SOA implementation3. Visibility refers to thecapacity of those with a need to see those witha capacity to service the needs. This isaccomplished by publishing the service in aregistry. Interaction is the act of using thecapacity. Typically, the interaction is performedthrough message exchange. The purpose ofusing the capacity is to get a real world effect.This effect may be a changed data state or a filetransferred, for example.

Technically, SOA is most often realized by usingWeb services. Web services follow the “Find,

Bind, Invoke” paradigm wherein the service thatneeds to be used is first found and then boundand invoked. The second and thethird steps of “binding andinvoking” are pure technicalproblems -- usually solved byusing Web Service DefinitionLanguage (WSDL) and SimpleObject Access Protocol (SOAP)messages. It is the first part of”finding” the service (which canbe theoretically solved by usingUniversal Description, Discovery, andIntegration, better known as UDDI), thatpresents the most interesting problem -- how dowe find the right service to be used?

Service Consumer

Bind andExecute

Service ProviderRegiste

r

Find

Contract Registry

The real world demands that systems have theability to intelligently make appropriateselection of services. An inference-drivenplatform for SOA can provide the ability to makeruntime decisions based on service capabilitiesand the context of invocation. SOA by itself doesnot have this capability. It is in this context thatsemantic richness can be of tremendous help8.Semantic richness can be thought of “as a wayof modeling a domain as close to naturallanguage as possible”. The primary goal of this

paper is to demonstrate how semantictechnology can be used to create an inference-driven platform for SOA.

The rest of this paper is organized as follows.The next section provides the background ofsemantic integration and ontology. After that,the paper discusses our experimental study ofsemantic technology and outlines the ontologymodel. We conclude by offering a direction forfuture research.

The real worlddemands that systems have theability to intelligentlymake appropriateselection of services.

white paper

Semantic Integration and Ontology

Semantic integration techniques attempt tobuild context or awareness about theunderlying services or data in the data sourcesin a meta-layer. Semantic Integrationtechniques are based on the axiom:Heterogeneous data (for a domain) may, onanalysis, reveal patterns of similarity in theirorigin, making it possible to define thesepatterns via a meta-model and to inter-relatethe meta-models by defining relationships.

Ontology-based approaches are considered asone way of achieving semantic integration.

Before we proceed to ontology-based SOA, abrief background of ontology is relevant. Anyarea of knowledge (domain) is usually expressedby splitting the domain to “classes” and byexpressing relationships between those“classes”. Ontology is defined as a “structuredrepresentation of a domain in terms of classesand interrelationships between the classes.” Forexample, if we want to understand projectmanagement, the ontological approach wouldbe1, 2:

3

Fig.2 Ontology chart for project management (fromLui, 2000, pp. 79)

Person

Society

Project

Assigned to

Task

Works on

Responsible forDepartment

#function

(employee)

(employer)

#hourly rate

#budget

#budget

#time spent

Organization

EmploysWorks

Project Management (Domain)

>Person (Class)

n Department (Individual)

>Organization (Class)

n Department (Individual)

n Project (Individual)

n Employer (Individual)

There are several relationships possiblebetween classes: person lives in a society,organization exists in a society, employee hasan employer, project has tasks, organizationdefines a project, employee works on a project,employee is assigned to a task, etc.

Ontology enabled SOA8 tries to extend and use thismodeling technique. Computer science requiresontology to be specified in the form of WebOntology Language (OWL) for processing. OWL isan XML- based specification. Fundamentally,ontology-enabled SOA resembles the following:

Briefly, in an ontology-enabled SOA, a layer iscreated that resides on top of the servicesdictionary that contains the ontology mappingsof the services domain. This is mapped withsufficient data during the design time which isthen used during the runtime to find semanticsimilarity of the service request. Policies arebuilt that contain logic for finding semanticsimilarity.

Selene: An Experimental Study

Selene is a Cognizant-designed architecturalframework created to understand the efficacy ofusing ontology to enable semantic integration. Therest of this section focuses on the description of theproblem domain that was addressed, thearchitectural solution and study results.

The Problem Domain

Drug discovery is an intensely collaborativeexercise wherein several scientists working invarious disciplines and systems are involved. Itis a highly data-driven process where hugevolumes of data are created and analyzed thatneeds to be shared with numerous scientists.The more the number of people and systemsinvolved in any activity, the greater theinformation integration challenges. This isespecially pronounced in an intensely R&D-based activity like drug discovery where varioustypes of applications are used by molecularbiologists, biochemists, toxicologists,pharmacologists, etc. The existing integrationtechniques that focus on data and their

drawbacks are dealt with more extensively in5

wherein the authors also lay a strong case forthe need for semantic techniques.

In the recent past, business realities havecompelled drug discovery toembrace a collaborative approachinvolving several partners. This hasopened up additional applicationintegration challenges, particularlyas the IT landscape has diversified.Initiatives like SIMDAT9 to enablecollaboration in the drug discoveryrealm have gained momentum. As aresult, many R&D organizations inthe pharmaceuticals business now use SOA asan effective medium for integration across theislands of information silos. Consider, forexample, an application that needs to be builtintegrating various data sources like:

n Chemistry

n Biology

n Materials

n Lab

n Assay Results

n Syndicated Data (Public Domain andSubscribed)

n Scientific Journals

One can create various atomic services fromthese underlying data sources using a technologyplatform that enables the composition ofapplications by linking services together.

white paper 4

Service Requests

Services

Ontology Mappings

Selene is a Cognizant-designed architecturalframework created tounderstand the efficacyof using ontology to enable semanticintegration.

white paper5

This paper focuses on a particular problemdomain within drug discovery area called assample management. All drug discovery activityneeds chemical and biological samples. Samplemanagement deals with the way the samplesare acquired, registered, stored and distributed.It can be defined as “activities in and aroundsample acquisition, registration and supply of

samples for drug discovery purposes.” Samplemanagement thus involves dealing with avariety of data sources for registration, storageand distribution of samples.

The figure below depicts typical areas that fallunder sample management.

Researchers

IntegrityAssurance

and ProcessQuality HTS

Sample Inventory

Registration

Sample Acquisition

Sample Management System

Chemists

Ordering

Scientists OperationsManager

Doctors Biologists

For the purpose of Selene, we chose an area ofsample management called sample acquisition.Sample acquisition deals with the act ofacquiring samples that are required for assaypurposes. The samples can be ordered by

scientists. The following table illustrates the“points of variability” where semanticintegration is required. Points of variability arethe areas where inference capability is requiredto have semantic richness.

Variability Matrix Request Vendor Selection Ordering Identification

Acquisition Process Raised by [Scientist,Branch, HTS, Ext.Research Org.]

Source [Chemists,Pharmaceutical Company,Commercial SampleLibrary]

Vendor [For {VendorSelection source}]

Type [Chemical, Biological]

Type [Chemical, Biological] Container Type [Vial, Plate] Type [Single, Bulk] Uniqueness Logic[SMILES, Sequence]

Name [Common, IUPAC] Sample State [Solid, Liquid] Container Type [Vial, Plate]

Molecular Formula Bar Code Sample State [Solid, Liquid]

Sample State [Solid, Liquid] Purity/Concentration Purity/Concentration

Container Type [Vial, Plate] Medium [Buffer, Vacuum,Salt Mixture, None]

Quantity [Volume, Weight] Location/Site

Purity/Concentration Bar Code

Quantity [Volume, Weight]

Price

As can be seen from the table, there are severalvariables for each step. It may be that thesevariables are today well defined and handledappropriately by existing systems. However,future business needs may mandateintroduction of additional variables or simplynew values for existing variables (e.g., a newcontainer type or ability to handle gaseous statesamples). Ideally we would like to have a servicedefinition, for example “Sample Registrationservice,” which should not require changes forincorporating the new variables. Similarly, theneed to incorporate a new vendor to supplysamples, for example, should not mandatechanges to existing acquisition service

implementation. Here is where defining anontology model for these variables enables anextensible vocabulary and a “non-invasive” wayto meet new business needs.

Solution Architecture

Our architectural solution, christened “Selene”was created as an SOA-based framework. Weused IBM products6 to develop this architecture.It contains workflow tasks and core businessprocesses designed as Service ComponentArchitecture (SCA) components with Javaimplementations.

white paper 6

Lab Technician Scientist

Vendor 1

Vendor 1OrderingService



BiologicsRegistration

Service

ChemicalRegistration

Service

Vendor 2 Vendor 3 Biological Chemical

RaiseSample

Requests

AcquisitionService

InventoryProcess

ContextSensitiveInvoker

SeleneOntology &Assertions

WebService

Endpoints

Web Services

BPEL

SelectVendor & Raise Order

Receive & Sample

Sample DispersalProcess

The architecture consists of the following broadcomponents:

n Business Process Modeler

n The Ontology Model

n Context Sensitive Invoker

n Web Services

Business Process Modeler

Business Process Modeler is used to model thebusiness process into discrete tasks. The“Sample Acquisition” business flow is modeledas follows:

white paper

Briefly, the workflow of sample acquisition is asfollows. A scientist requests a sample. Thissample may either exist in the inventory, mayhave been in inventory at some time but is nowdepleted, or may be a completely new sample(i.e., never had an entry in the inventory). Thepharmaceutical company may have a list ofpreferred vendors for supplying samples; thescientist can chose to place the order with oneof these vendors. After the sample is ordered,when it is received, depending on whether thesample existed earlier in inventory or not,registration of the sample is required. Once thesample is registered, it is available for dispersal.This flow includes three main human tasks:

n Raise Sample Request

n Choose Preferred Vendor to Order

n Receive Sample

Once the process is defined, it is expressed inBusiness Process Execution Language (BPEL)form in the Business Modeler.

The Ontology Model

As described in the introduction, Ontology isdefined in terms “concepts” and “relationships.”Concepts are implemented in ‘classes” while“relationships” are implemented in terms of“Object Properties” and “Data type properties”4.

For the sample acquisition process that isdescribed in the previous section, we define aconceptual ontology model. The following arethe main classes in our model:

n Sample

n Sample Request

n Sample Order

n Scientist

n Lab Staff

n Vendor

For the purpose of simplicity, in our ontology wehave chosen to represent the below subset ofthe classes and their relationships, based on theaforementioned variability matrix. Theseclasses are grouped under appropriately namedsuper classes as shown in the following table.

7

SampleInventory

Check

RetrieveVendor List Choose

PreferredVendor

ReceiveOrderedSample

PersistOrder

OrderSample toPreferredVendor

IdentifySample

PrintInventory

Label

InventoryProcess

SampleDispersalProcess

SampleRequest

50.0% No

50.0% YesIs SampleAvailable?

Sample Request

Sample Request

Sample Request

Sample Request

Vendor List Sample Order

SampleRequest

Sample Request

Sample Request

Sample Request

Sample Request

Sample Request

Sample Inventory Panel

Sample OrderSampleOrder

white paper

The Actor class is a suitable abstraction ofentities like Vendor and Scientist. Similarly aSample Request or Sample Order is a kind ofDocument while Material appropriatelyabstracts an entity like Sample.

Similarly we use Attribute class to representattributes and their possible values (for

example, “Sample Type” is an Attribute of theMaterial, “Sample,” can have a value of“Chemical” or “Biological.”)

The above definitions and classifications allowfor the visualization of the classes and theirrelationships in a generic manner conforming toontology representations as shown below:

8

Relationship

Object Property

ClassClass

Actor/Document/Material/Attribute

Actor/Document/Material/Attribute

Or more specifically:

Super Class Class Relationship Attributes

Material Sample hasNamehasSampleTypehasStatehasContainerTypehasStorageTemperaturehasStructure

NameSample TypeSample StateContainer TypeStorage TemperatureStructure

Document Sample Request containshasRequestedQuantityhasRequester

SampleRequested QuantityRequester (Lab staff)

Document Sample Order containscontains

VendorSample Request

Actor Scientist isEmployedWith Organization

Actor Lab Staff isEmployedWithreports to

OrganizationScientist

Actor Vendor hasNamehasRatinghasVendorType

NameRatingVendorType

white paper

The diagram below helps to visualize ourontology model for the sample acquisitionprocess.

The runtime ontology is created using RationalSoftware Architect 7.5. Currently, Selene uses“Vendor Name” and “Sample Type” assertionsto define end point selection policies. Other

For our ontology, we could represent therelationship between the Actor (Vendor) andthe Attribute (Name) with the Object Property“hasName.” The same property can also depictthe relationship between the Material (Sample)and Attribute (Name). In other words, we can

say that the domain for the object property“hasName,” includes the Actor (Vendor), andthe Material (Sample), while the Attribute(Name) is in its range. This can be expressed inOWL as below:

9

<owl:ObjectProperty rdf:ID="hasName"><rdfs:range rdf:resource="#Name"/><rdf:type

rdf:resource="http://www.w3.org/2002/07/owl#FunctionalProperty"/><rdfs:domain>

<owl:Class><owl:unionOf rdf:parseType="Collection">

<owl:Class rdf:about="#Sample"/><owl:Class rdf:about="#Vendor"/>

</owl:unionOf></owl:Class>

</rdfs:domain></owl:ObjectProperty>

ctssm:Material

ctssm:Sample

ctssm:SampleState

ctssm:SampleType

ctssm:ContainerType

ctssm:VendorType

ctssm:Structure

ctssm:Type

ctssm:Rating

ctssm:Vendor

ctssm:Scientist

ctssm:Actorctssm:SampleRequest

ctssm:LabStaff

ctssm:Name

ctssm:Organization

ctssm:Storage

Temperature

ctssm:RequestedQuantity

ctssm:hasName

ctssm:hasContainer

Type

ctssm:hasSample

Type

ctssm:hasState

ctssm:hasStorage

Temperature

ctssm:hasRequestedQuantity

ctssm:hasRating

ctssm:hasStructure

ctssm:containsSample

ctssm:containsSampleRequest

ctssm:containsVendor

ctssm:hasName

ctssm:isEmployed

With

ctssm:isEmployed

With

ctssm:Reportsto

ctssm:hasRequestor

ctssm:hasRequestor

ctssm:hasVendorType

ctssm:hasName

isa

isa isaisa isa isa

isaisa isa isa

isa

isa

isaisa

isaisa

ctssm:Attribute

ctssm:SampleOrder

ctssm:Document

white paper

with a new supplier for samples. Selene wouldneed no change to its core business process asimplemented in BPEL. Extending the ontologyfor the new vendor and defining the new endpoint with a policy assertion binding wouldsuffice. Of course, the new vendor would needto expose its business interface as a Web serviceeither registered on a UDDI registry or share alibrary of services expressed in WSDL.

Context Sensitive Invoker

The business flow is encapsulated as a linearprocess in a BPEL block. The “Context SensitiveInvoker” uses ontology mapping to determinethe appropriate Web service. This is done by“Assertions” and “Policies”. Policies operate onthe available and published end points.Assertions are the ontology mappings of allpossible context-sensitive variables. Policies aredefined in terms of these assertions -- usually asubset of assertions. Context Sensitive Invokerdetermines at runtime the end point to invokebased on policy assertions.

Web Services

Web services provide the actual businessservice. For Selene these are considered as“consuming” points in the sense that Webservices are invoked. From an architecturalframework perspective, theoperational assumption is thatWeb services will be madeavailable for use.

The policy to be defined on theWeb services is very much part ofthe tasks required to incorporatea new service endpoint.

Study Results

The diagram on the following page illustrateshow the Selene architecture works at runtime.

A BPEL process is initiated on receipt of a“Sample Request” which contains samplecriteria like sample state, preferred vendor, etc.

10

assertions are included for possible future use.The runtime ontology mapping as representedin the Unified Modeling Language (UML) form isshown below:

Selene uses IBM Fabric 6.2 as the core ontologyengine. The ontology, as implemented in Selene,achieves portability and scalability throughloose coupling. As an example, apharmaceutical company can forge partnership

SampleType Enumeration

SampleType Assertion

01-sampleTypeEnumeration

01-samplePhysicalStateEnumeration

01-sampleContainerType Enumeration

<<owlClass>>

<<fabricAssertion,owlClass, rdfs:label>>

hasSampleType:string

SamplePhysicalStateAssertion


hasPhysicalState:string

SampleContainerTypeAssertion


hasContainerType:string

VendorNameAssertion


hasVendorName:string

VendorRatingAssertion


hasVendorRating:string

<<fabricEnumeratedProperty>>

hasSampleType


hasPhysicalState


hasContainerType

01-vendorNameEnumeration


hasVendorName

01-vendorRatingEnumeration


hasVendorRating

SampleContainerType Enumeration

<<owlClass>>VendorName Enumeration

<<owlClass>>VendorRating Enumeration

<<owlClass>><<owlClass>>SamplePhysicalState Enumeration

<<owlClass>>ContentBasedAssertion

The ontology, asimplemented inSelene, achievesportability and scalability throughloose coupling.

white paper

The BPEL engine of the “Business ProcessModeler” passes this on to the “ContextSensitive Invoker.” The SCA components ofContext Sensitive Invoker interact with the“Ontology Engine” and retrieve the assertionsbased on the context which in this illustratedcase is “Vendor.”

The “Vendor Name Assertions”provide a reference to match theinput “Vendor” information withthe policy assertion bound to theend points. The required WSDL toexpose the ‘Vendor’ web services isobtained by generating a webservice binding with SOAP/HTTPtransport with JAX-RPC headers forthe respective SCA imports.

In Selene, we demonstrate the proper “VendorService” end point selection by simply loggingthe name of the selected end point. Similar tothe vendor assertions, “Sample” related policyassertions (e.g.. Sample Type) assist the ContextSensitive Invoker to hit either the BiologicsRegistration or Chemical Registration Webservice based on received Sample Type.

Thus the desired context-sensitive selection ofWeb services is gained using ontological means.We further note that it is relatively simple toachieve this using ontology as the medium forsemantic integration.

Conclusions & Future Work

In this work, we have shown that it is possible touse ontology for enabling semantic integration.The approach followed was to model the businessdomain process as a workflow. Then the points ofvariability were identified for each of the workflownodes and the decision points. The ontologymodel construction was a parallel activity thathappened alongside the business processmodeling. The classes and relationships of theontology model are created based on the domainknowledge and the business process model. Thisontology was then converted to Web ontologylanguage form for computational purposes. Thecontext-sensitive invoker, which is the core of thisstudy, determines the runtime invocations basedon the ontology model. Thus, we have achievedsemantic integration using ontology.

11

SelectVendor &

Raise Order

BPEL

Preferred Vendor=Vendor 2

Web Service Endpoints

Matching PolicyAssertion

Vendor Web Services

SeleneOntology

Context Sensitive

Vendor 1 Ordering Service



Vendor 1 Vendor 2 Vendor 3

Further work needs to be done to create a

more complete ontologythat encompasses not

just sample acquisitionactivity but the entire

sample managementoperations.

The ontology that is shown in this paper is notcomplete. Further work needs to be done to createa more complete ontology that encompasses notjust sample acquisition activity but the entiresample management operations. We plan tostrengthen the domain ontology model by addingmore domain specific information. Of relevance, isthe SOA-based ontology model proposed by TheOpen Group10. We intend to create an ontologymodel based on the approach defined by TheOpen Group.

From a technical perspective, we feel thatontology is a powerful approach for creatingintelligent software agents. Work on usingontology for query reformulation represents aninteresting possibility for use of ontology11. Suchperceptive ability is also required in softwareagents to make runtime decisions; we believethat it can be built using ontology. This can helpin creating user friendly software that helps in

decision making. For example, in the domain ofdrug discovery, we can create a software agentthat using ontology analyzes the samplerequests of the scientists and suggests potentialalternatives across the vendorcatalogues. We plan toinvestigate this approach.

Acknowledgements

The authors thank IBM for theirsupport in realizing the Selenearchitecture. We thank RakeshRonald, Business Analyst atCognizant, for his timely andvaluable help in modeling the business process.We express our gratitude to BalasubramanianSankaranarayanan, Practice Director for LifeSciences at Cognizant, for his encouragementand guidance in this effort.

white paper 12

Further work needs to bedone to create a morecomplete ontology thatencompasses not justsample acquisition activitybut the entire samplemanagement operations.

About the Authors

Raghuraman Krishnamurthy is a Principal Architect in Cognizant’s Technology Consulting Group,working within the life sciences domain. His focus spans software development and linking IT tobusiness and software architectures. Raghu’s technical area of expertise includes enterprisearchitecture, SOA and application integration. He has been involved in various consultingengagements in architecture assessment, application integration. Raghu holds a Master’s degreefrom the Indian Institute of Technology, Mumbai. He is a TOGAF certified enterprise architect.

Vinod Ranganathan works as a Technical Lead on J2EE/SOA platform projects for variouspharmaceutical clients of Cognizant. He has considerable experience in the areas of applicationdevelopment, project management and technical consulting. Vinod is also involved in design andexecution of PoCs in addition to architecture recommendations and software estimation forpotential projects. He holds a Masters degree in Biotechnology from Pune University and a Diplomain Advanced Computing from C-DAC. He is also working towards a Post-Graduate Diploma inBusiness Administration from Symbiosis Institute.

Baskar Senguttuvan is a Senior Architect in Cognizant’s Technology Consulting Group workingwithin the life sciences domain. He has expertise in architecture consulting and projectmanagement. Baskar has consulted with various pharmaceutical clients in the areas of SOA andinformation technology architecture, among other things. He holds Bachelor’s Degree in ComputerScience and MBA in Systems. He is certified PMP and SEI Software Architecture Professional fromCarnegie Mellon’s Software Engineering Institute.

World Headquarters

500 Frank W. Burr Blvd.Teaneck, NJ 07666 USAPhone: +1 201 801 0233Fax: +1 201 801 0243Toll Free: +1 888 937 3277Email: [email protected]

© Copyright 2009, Cognizant. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, transmitted in any form or by any means, electronic, mechanical, photocopying, recording, orotherwise, without the express written permission from Cognizant. The information contained herein is subject to change without notice. All other trademarks mentioned herein are the property of their respective owners.

India Operations Headquarters

#5/535, Old Mahabalipuram RoadOkkiyam Pettai, ThoraipakkamChennai, 600 096 IndiaPhone: +91 (0) 44 4209 6000Fax: +91 (0) 44 4209 6060Email: [email protected]

European Headquarters

Haymarket House28-29 HaymarketLondon SW1Y 4SP UKPhone: +44 (0) 20 7321 4888Fax: +44 (0) 20 7321 4890Email: [email protected]

About Cognizant

Cognizant (NASDAQ: CTSH) is a leading provider of information technology, consulting and businessprocess outsourcing services. Cognizant’s single-minded passion is to dedicate our global technology andinnovation know-how, our industry expertise and worldwide resources to working together with clients tomake their businesses stronger. With over 50 global delivery centers and 63,700 employees as of March31, 2009, we combine a unique onsite/offshore delivery model infused by a distinct culture of customer sat-isfaction. A member of the NASDAQ-100 Index and S&P 500 Index, Cognizant is a Forbes Global 2000 com-pany and a member of the Fortune 1000 and is ranked among the top information technology companiesin BusinessWeek’s Hot Growth and Top 50 Performers listings.

Start Today

For more information on how to drive your business results with Cognizant, contact us [email protected] or visit our website at www.cognizant.com.

References

1. Rodrigo Bonacin, M Cecilia C. Baranauskas, Kecheng Liu 2004 FROM ONTOLOGY CHARTS TOCLASS DIAGRAMS: semantic analysis aiding systems design Proceedings of the 6th InternationalConference on Enterprise Information Systems, Porto, 2004.

2. Liu, K., 2000. Semiotics in information systems engineering, Cambridge University Press.

3. Wikepidia on Semantic Integration, Ontology and Service Oriented Architectures

4. Protégé ontology development tools, http://protege.stanford.edu/download/ontologies.html

5. Slater, T, Beyond data integration, Drug Discov Today (2008), doi:10.1016/j.drudis.2008.01.008

6. IBM literature on SOA and tools

7. Mike P. Papazoglou, Willem-Jan van den Heuvel, Service oriented architectures: approaches,technologies and research issues The VLDB Journal — The International Journal on Very LargeData Bases, Volume 16 , Issue 3 (July 2007), Pages: 389 - 415

8. Semantics to Energize the Full Services Spectrum, Amit Sheth, Kunal Verma, Karthik Gomadam,July 2006, Communications of the ACM, Volume 49, Issue 7

1. SIMDAT http://www.simdat.eu/859.html

2. The Open Group, SOA domain ontology http://www.opengroup.org/projects/soa-ontology/uploads/40/16940/soa-ontology-200-draft.pdf

3. Nwe Ni Tun Dec 2006, Semantic Enrichment in Ontologies for Matching, AOW '06: Proceedings ofthe second Australasian workshop on Advances in ontologies - Volume 72

http://www.opengroup.org/projects/soa-ontology/uploads/40/16940/soa-ontology-200-draft.pdf

http://www.opengroup.org/projects/soa-ontology/uploads/40/16940/soa-ontology-200-draft.pdf

http://www.simdat.eu/859.html

Drug Discovery: An Innovative Approach to Application ... · compelled drug discovery to embrace a...

Documents

Transcript of Drug Discovery: An Innovative Approach to Application ... · compelled drug discovery to embrace a...