Department of Biomedical Informatics Development of Ontology-anchored Grid-based Data Services to Facilitate Integrative Clinical and Translational Science

Department of Biomedical Informatics

Development of Ontology-anchored Grid-based Data Services to Facilitate Integrative Clinical and Translational

ScienceIntroduction

Rakesh Dhaval

The Ohio State University Medical Center


Topics

• Use Case• TRIAD• openMDR• TRIAD – openMDR Workflow• Questions


Use Case

A clinical researcher is conducting an outcomes study in which they want to utilize data collected during the course of standard clinical care, as well as bio-specimen data from the Tissue Bank.

For patients that have consented to have their tissue specimen stored in the Tissue Bank with specific sexGenotype, ethnicity and gender - The researcher is interested in knowing Pre-Operating Room Diagnosis and Procedure, as well as Post-Operating Room Diagnosis and Procedure.

The researcher would like to be able integrate and analyze all of this information as a single data set.


Bedside-Bench

Bench-Bedside

Basic Sciences

Hypothesis Discovery

Analysis & Dissemination

Pre-Clinical Research

Clinical Practice

Observation &

Surveillance

Clinical Research

Hypothesis Formulation

Translational Research

Translational Research


Topics



The Translational Research Informatics and Data Management Grid (TRIAD)

TRI D


TRIAD Introduction

• TRIAD = CTSA + caGrid Data Service Wizards Federated and Localized Metadata Management

• caBIG - Cancer Biomedical Informatics Grid• Grid - “Controlled and coordinated resource sharing and problem solving in

dynamic, scalable virtual organizations.”• caGrid - Provides the software infrastructure that underlies the tools and

applications of caBIG

TRIAD middleware


• TRIAD is an instance of caGrid that leverages existing, core components: Service oriented architecture utilizing Axis, Globus, and caGrid extensions. Data model management via the Global Model Exchange. Customizable security infrastructure leveraging GAARDS. Graphical strongly typed grid service creation using Introduce. Extendable and customizable web service infrastructure.

• TRIAD extends caGrid: Additional options for data service creation, security policy, metadata management, etc. Federated and/or local metadata management utilized for semantically annotated grid

service creation “Out of the box” data services for popular tools such as I2B2 and RedCAP Integration with commercial and open-source CTMS platforms

TRIAD Design

TRI D


Compatibility and Interoperability

• caBIG® provides standards-based compatibility guidelines for creating software systems that are syntactically and semantically interoperable.


Knowledge Engineering Framework

Borlawsky TB, et al. Development of an Agile Knowledge Engineering Framework in Support of Multi-Disciplinary Translational Research. Proc AMIA Translational Bioinformatics Summit, 2009.


Example Use Case


Use Case: Technical Implementation


Enhanced Data Services

• Data service generation wizards for common data centric applications

Class Model & Assertions

CDEs & Semantics


Create Semantically Harmonized Data Model Generate Data

Resource

Grid-ify

TRIAD Development Process

Create an Information Model using a modeling

tool

Information Models

Perform Semantic

Integration using the EA Plugin

Vocabularies

Generate Code and Interfaces using the caCORE SDK

Code GeneratorAPIs

Transform the Model

into Metadata using the

Domain Model Gen.CDEs

Generate a Grid Service

using Introduce

Grid

Reference: Dr. Robert Freimuth, Vocabulary Knowledge Center Director

• caCORE is a robust set of tools and resources to support the development of caBIG®-compatible systems

• NCI offers comprehensive training for caCORE tools


Topics



OpenMDR-based Federated WorkFlow

openMDR - TRIAD based domain modeling and annotation workflow


openMDR: Components


• Metadata Registry - Capable of storing, versioning, and maintaining semantic and representational metadata.

• Utilizing/enhancing the UK cancergrid cgMDR ISO11179 database.

• Built on eXist database.» Metadata is annotated

with conceptual information from a local or remote lexEVS system(s).

» Web based front-end for creating and curating semantic metadata.

openMDR:mdrCORE


• API and Grid Service for querying across many disparate semantic metadata repositories» caGrid grid service - capable of talking to caDSR and many openMDR systems enabling federated query of common data elements.

» Enhances and wraps the cgMDR mdrConnector in order to parse the received information into a common format.

» caGrid service enables it to be deployed in any local or production grid environment and can be configured and modified to talk to many semantic metadata systems.

openMDR:mdrQuery


• Semantic Annotation Tool» Enables semantic annotation of federated semantics

» Utilizes MDRQuery service for locating and utilizing CDE’s from multiple semantic metadata sources

» Options for Search Filters» Enables Search within “Context”» Tags are created that identify CDE via local unique id

» No need to leave the EA environment

openMDR:mdrEAPlugin


• Process an MDR Plugin annotated model and create service metadata which will enable the service to be a semantically annotated grid service and easily discovered» The metadata can be used to generate a data service using the Introduce Data Service Wizard.

openMDR:mdrDomainModelGenerator


Topics

• Use Case• TRIAD• openMDR• TRIAD – openMDR WorkFlow• Questions


TRIAD - openMDR Based Service Creation Workflow

openMDR - TRIAD based Service Creation workflow


① CDE curation in Local instance of openMDR - using concepts from Local or Remote (NCI) LexEVS instance

② Create an information model - create logical model, data model and Map Object Model to Data Model in Enterprise Architect (EA)

③ Annotate logical model with CDE's in EA - using openMDR EA plugin.

④ Export annotated XMI

⑤ Create domain model file (XML file) with grid service metadata

⑥ Generate a system using caCORE SDK 4.X by supplying your UML model as input

⑦ Create Grid Service - Use Introduce grid Service Authoring Toolkit

OpenMDR Service Creation Phases


Topics

• Use Case• TRIAD• openMDR• TRIAD – openMDR WorkFlow• Questions


Questions/Comments?

Thank you for your time and attention

TRIAD: Dave Ervin [email protected]

http://wiki.triadcommunity.org

openMDR: Rakesh Dhaval [email protected]

http://cagrid.org/display/mdr

caGrid Knowledge Center: [email protected]

mailto:[email protected]

http://wiki.triadcommunity.org/





Development of Ontology-anchored Grid-based Data Services to Facilitate Integrative Clinical and Translational

Science

TRIAD: Walkthrough and Demonstration

Rakesh Dhaval

Applications Architect, The Ohio State University Medical Center


Topics

• Review of Use Case• Model Development in Enterprise Architect

– Data Model Development– Logical Model Development– Object Relational Mapping

• Semantic Annotation– Create Data Elements in openMDR– Annotate Model with Data Elements

• Acknowledgements• Questions


Use Case

A clinical researcher is conducting an outcomes study in which they are utilizing data collected during the course of standard clinical care, as well as bio-specimen data from the Tissue Bank.

For patients that have consented to have their tissue specimen stored in the Tissue Bank with specific sexGenotype, ethnicity and gender. The researcher is interested in knowing Pre-Operating Room Diagnosis and Procedure, as well as Post-Operating Room Diagnosis and Procedure.

The researcher would like to be able integrate and analyze all of this information as a single data set.


Topics

• Review of Use Case• UML Model Development



• TRIAD Portal • Acknowledgements• Questions


UML Model Development

Create a Logical

Model (UML class

diagram) using

Enterprise ArchitectLogical Model

Create a Data Model (database schema) using

Enterprise Architect

Data Model

Semantically Annotate the UML Model using the openMDR EA Plugin

Semantics

Map the Logical

Model to the Data Model

using caAdapter

Mapping

Model is complete and ready for review and load into caDSR/ openMDR

Load Model


Import Data Model

• Project > Database Engineering > Import DB Schema from ODBC ...


Data Model



Create a Logical

Model (UML class

diagram) using




Data Model


Semantics

Map the Logical


using caAdapter

Mapping

Model is complete and ready for

compatibility review and load into caDSR/ openMDR

Load Model


Logical Model : Add a Class

• Select Class• Click pointer in diagram workspace


Logical Model : Enter Class Name


Logical Model :Add Class Attributes

• Select ‘Details’ tab and click ‘Attributes’ button


Logical Model : Add Class Attributes

• Enter ‘Name’• Select Java ‘Type’

– lang• Boolean• Character• Double• Float• Integer• Long• String

– util• Date

• Click ‘Save’


Logical Model :Add Association

• Select ‘Associate’ tool from Class Tools menu

• Drag from Source to Target class

Source Target


Logical Model :Association Properties

• Double-click association• Specify directionality under

General tab• Specify multiplicity under

Source and Target tabs


Logical Model



Create a Logical

Model (UML class

diagram) using




Data Model


Semantics

Map the Logical


using caAdapter

Mapping


Load Model


ORM : Export XMI

• Right-click on Logical View• Select Import/Export > Export Package to XMI file...


ORM : Export XMI


ORM : caAdapter

• Select File > New > Model Mapping Service > Object Model to Data Model Map Specification


ORM : caAdapter

• Open the XMI file that was exported from EA


ORM : caAdapter

• Drag and drop to associate objects and tables, attributes and columns


Topics







Create a Logical

Model (UML class

diagram) using




Data Model


Semantics

Map the Logical


using caAdapter

Mapping


Load Model


MetaData:Common Data Elements


Topics






Semantic Annotation

• Creating New Data Elements using openMDR Metadata Registry

Demo

http://mdr-demo.triadcommunity.org:8080/


Topics






Semantic Annotation

• Annotate the Clinical Model in Enterprise Architect Tool

Demo


Semantic Annotations

• New concept created using openMDR– Logical model:

• Object: Visit• Attribute: dischargeDisposition

– Data element: Discharge Disposition cagrid.org_0edff23f-c0fa-4a18-ac06-97e4893d42e8_0.1

• caTissue link– Logical model:

• Object: Patient• Attribute: studyNumber

– Data element: Participant Identifier java.lang.Long (public ID: 2513658)• Other

– Logical model:• Object: Patient• Attribute: Gender

– Data element: Participant Gender Person Gender Text Type (public ID: 2513661)

https://140.254.126.108:8443/exist/rest/db/mdr/web/data_element.xquery?compound_id=cagrid.org_0edff23f-c0fa-4a18-ac06-97e4893d42e8_0.1


Topics






TRIAD Portal Demo

• User Interface for Investigator to Query DataTRIAD Portal – Under Development

Demo


Topics






Acknowledgements

• CTSA: http://ctsaweb.org/ • NCI caBIG

– caGrid toolset: https://cabig.nci.nih.gov/index_html/sharable/cagrid_overview

– caGrid Knowledge Center: https://cabig-kc.nci.nih.gov/CaGrid/KC/index.php/Main_Page

• Active caGrid Community

http://ctsaweb.org/

https://cabig.nci.nih.gov/index_html/sharable/cagrid_overview

https://cabig.nci.nih.gov/index_html/sharable/cagrid_overview

https://cabig-kc.nci.nih.gov/CaGrid/KC/index.php/Main_Page

https://cabig-kc.nci.nih.gov/CaGrid/KC/index.php/Main_Page


Topics






Questions/Comments?

Thank you for your time and attention

TRIAD: Dave Ervin [email protected]

http://wiki.triadcommunity.org

openMDR: Rakesh Dhaval [email protected]


caGrid Knowledge Center: [email protected]


http://wiki.triadcommunity.org/



Documents

Department of Biomedical Informatics Development of Ontology-anchored Grid-based Data Services to Facilitate Integrative Clinical and Translational Science