Towards a Set of Unified NIH Computational, Data, and Community Infrastructures to Support Translational Bioinformatics

1
Towards a Set of Unified NIH Computational, Data,
and Community Infrastructures to Support
Translational Bioinformatics

• Brian D. Athey, University of Michigan
• Mark Ellisman, University of California, San
  Diego
• Michael J. Becich, University of Pittsburgh
  Medical Center (UPMC)
• Joel Saltz, Ohio State University
• March 11, 2008
• 1st AMIA Summit on Translational Bioinformatics
• March 26, 2008
• CASC

2
We cannot nor should not underestimate the
future relationship between bioinformatics and
the omics and the future EHR

• Inclusion of modern genome-wide records into
  the Electronic Health Record (HER) was the number
  one request of HHS Secretary Leavitt to the NIH
  leadership at a small private dinner recently.
• -Donald AB Lindberg,
• Director, National Library of Medicine
• DoD Healthcare IT Summit
• March 26, 2008

3
Summary of NIH Panel Members Informatics Efforts

• NCBCsBasic science to translational informatics
  specialty centers. Roadmap (i.e. transitional).
  Athey
• BIRNFocused on community building, strong in
  neurosciences. NIH Cyberinfrastructure standard
  bearer. Ellisman/Saltz
• CTSAsRaised level of awareness of pent-up need
  for clinical and basic research informatics and
  IT and requirements to interoperate with
  in-patient and ambulatory IT systems.
  Becich/Athey
• caBIGFocused on operationalizing network of
  NCI-funded Comprehensive Cancer Centers. Saltz,
  Becich, Athey

4
Elephants in the Room

• Underspecified/non-existent research IT systems,
  architecture, and integration
• Cancer Center Directors who dont get caBIG or
  who dont want to
• R01 investigators who arent primarily interested
  in data sharing or in using somebody elses
  tools
• Lack of data sharing policies for basic and
  clinical
• Dichotomy of proprietary systems (Hospital/Health
  System) and open source (Research) solutions and
  the risk averse nature of Academic Health Centers
• Think interfaces
• Some research systems proprietary (e.g. Velos,
  ONCOR, etc.)
• Hospital/In-patient and Ambulatory Care
  Information Systems (ACIS) interoperability with
  research IT systems

5
National Infrastructure to Leverage

• Clinical and Translational Sciences Award (CTSA)
  Informatics Consortium (NCRR)
• 24 Members to grow to 60 in 3 years
• NIH National Biomedical Computing Centers (NCBCs)
• Biomedical Informatics Research Network (BIRN
  NCRR). Related NIH Cyberinfrastructure efforts
  (CDI, DATANet, etc.)
• caBIG

6
Panel Goals

• Learn more about these key NIH Informatics
  Infrastructure programs
• Begin a dialog, from the bottom-up, to identify
  key commonalities and synergies possible between
  these programs.
• Answer the question what has to change??

7
How can we build upon and sustain these efforts?

• Its more than standardization, ontologies, and
  harmonization
• It is about engaging individual investigators
  with team science
• How do we bridge this cultural divide?
• How do we balance an individual and individual
  institutions needs with national needs?
• E.g. pooling of genomics data to build the number
  of subjects for statistical power.
• We might be preaching to the choir here. Please
  spread the word back home and in DC.

8
NCBC PortalLook under the buttons
www.ncbcs.org
9
iTools Prototype
10
NCBC Categorization of Scientific Ontologies
11
Categorization of Scientific Ontologies
Domain Prefix Category OBO Foundry
Biological process GO 1 All NCBCs endorse yes
Cell type CL 3 Promising but under construction yes
Cellular component GO 1 All NCBCs endorse yes
Chemical entities of biological interest CHEBI 3 Promising but under construction yes
Current Procedural Terminology CPT 2 All NCBCs will use under protest (or more often, with a wish for some additions/corrections) no
FlyBase FB 1 All NCBCs endorse no
12
Scope of Applications in CTSA Informatics

• Interoperability with Institutional EMR Systems
• Clinical transaction systems
• Clinical Data Repository (CDR)
• De-identification/Honest Brokering
• Tools to Facilitate Extracting/Downloading Data
  Software tools
• CTSI Portals
• Clinical Trial/Study Databases
• Genomic, Proteomic, and Metabolomic
  High-Throughput Data Repositories and Analysis
  Tools
• Clinical Imaging Data Repositories and Analysis
  Tools
• An Institutional Specimen Tracking System
• A CTSA Core Lab LIMS (Laboratory Information
  Management System)
• Population/Public Health Databases Informatics
  Needs
• Standards to promote Interoperation within and
  between CTSA sites
• Informatics Teaching Training (Interface with
  CTSA Education Program)
• Biomedical Informatics Research in Support of CT
  Research
• Faculty, Staff, and Administrative Structure for
  Biomedical Informatics

13
CTSA Consortium

• CTSA National Informatics Steering Committee
• Project Incubator
• Data Sharing (Lead - TBD, Liaison - Silverstein)
• CTSA Informatics All-Hands Meeting before AMIA
  (Lead - TBD Liaison - Masys)

CTSA Informatics Operations Committee

• Interest Groups
• Collaboration Facilitation
• Education
• Data Warehousing
• User Needs
• Standards Interoperability

• Project Groups (tentative)
• CTSA Informatics Priorities (Leads - Athey,
  Miller)
• Clinical Research Registry (Lead - Sim,
  Liaison - Silverstein)
• Education (Lead - Klee, Liaison - Hersh)
• Inventory (Lead - McWeeney, Liaison - Becich)
• IT/Informatics White Paper (Nearing
  completion)

Interest Groups Propose Projects
Interest Groups Propose Projects
14
caBIG

• Must get smaller and scale to the user
• Must get bigger by scaling to the enterprise
• Must normalize with NIH CTSA Informatics,
  NCBCs, and BIRN.
• NCI must continue to invest in the CCCs
  personnel to adapt caBIG
• Cancer Center Directors need the Fear of God
  relating to non-adaption.

15
caBIG Community
Source caBIG Ready for Adoption/Adaption. Ken
Buetow, Ph.D., June 21, 2007
16
caBIG November 8-9, 2007 Summit
Recommendations Summit participants, following a
day and a half of deliberations in three
simultaneous subject tracks, achieved consensus
on the following eight priorities for the caBIG
initiative 1. Sustain its work in data standards
and infrastructure. 2. Spearhead an awareness
campaign. 3. Conduct a scientific demonstration
project. 4. Maximize engagement with the
commercial sector. 5. Establish more extensive
and visible partnerships with other government
agencies. 6. Get inside Electronic Health
Records. 7. Expand beyond cancer. 8. Expand
internationally.
Source caBIG Summit Executive Summary (January
2008)
17
Source Onsemble Notes Newsletter of the
Oncore Community, Vol 3., No 1. (Spring 2008)
18
CTSA Institutions Oncore / Velos

• Funded CTSA Institutions

• 2006 Awardees
• Duke University (Velos)
• Columbia University (Velos)
• Mayo Clinic
• Oregon Health and Science Univ (Velos)
• The Rockefeller University
• Univ of California, San Francisco (Velos)
• Univ of California, Davis (Velos)
• University of Pennsylvania
• University of Pittsburgh
• U Rochester Sch of Medicine and Dentistry
• U Texas Health Sciences Center at Houston
• Yale University

• 2007 Awardees
• Emory Univ (with Morehouse) (Oncore)
• CWRU / Cleveland Clinic (Velos)
• Weill Cornell Medical College (with Hunter)
• Johns Hopkins University
• Univ Of Michigan At Ann Arbor (Velos)
• U Texas Southwestern Med Ctr - Dallas
• Univ Of Wisconsin Madison (Oncore)
• University Of Chicago (Velos)
• University Of Iowa (Oncore)
• University Of Washington
• Vanderbilt Univ (with Meharry) (Oncore)
• Washington University

http//www.ctsaweb.org
19
Key Challenges and Opportunities

• Sociological and Technical Complexity of the
  Informatics and IT environments
• Must SIMPLIFY for users to use.
• Must continuously educate our users
• We must deploy proprietary and open source tools
  with the larger integration picture in mind.
• We must build and leverage specialized and
  standardized IT production resources in our AHCs.
• NIH can help us leverage these national
  initiatives with attractive supplemental programs
  which insist on using tools from these
  initiatives.
• Cancer Center leadership must make Clinical
  Research Informatics a top priority (it is a
  Force Multiplier)
• A co-investment strategy with NIH and the AHCs
  and other willing partners (public and private)
  is called for.
• Lets sustain this discussion at AMIA, engaging
  our willing NIH colleagues. Lets start now.