NPACI Neuroscience

1
Molecules
Synchrotrons
Macromolecular Complexes, Organelles, Cells
Microscopes
Organs, Organ Systems, Organisms
Magnetic Resonance Imagers
2
Biomedical Informatics Research Network a testbed
for a biomedical knowledge infrastructure
3
BIRN Project Objectives
1) establish a stable high performance network
linking key Biotechnology Centers and General
Clinical Research Centers 2) establish
distributed and linked data collections with
partnering groups 3) enable the use of
"grid-based" computing 4) enable data mining
from multiple data collections or databases on
neuroimaging 5) build a stable software and
hardware infrastructure that will allow centers
to coordinate efforts to accumulate larger
studies than can be carried out at one site.
4
EACH BRAIN REPRESENTS A LOT OF DATA
AND COMPARISONS MUST BE MADE BETWEEN MANY
Slide courtesy of Arthur Toga / UCLA
5

6
Federate Emerging Databases
Infrastructure to relate, combine produce
meta data
Deformation Segmentation Quantification
Slide courtesy of Arthur Toga
7
(No Transcript)
8
Two groups have developed BIRN project
partnerships, so far

• Mouse BIRN - Animal Models of Disease / Multi
  Scale/Multi Method - MS Mouse and DAT KOM (a
  schizophrenic and otherwise interesting mouse
  animal model)
• Brain Morphology BIRN - Targets neuroanatomical
  correlates of neuropsychiatric illness (Unipolar
  Depression, mild Alzheimer's Disease (AD), mild
  cognitive impairment (MCI)

9
THE BIOMEDICAL INFORMATICS RESEARCH NETWORK
LAYOUT FOR PHASE 1
10
Getting down to work - BIRN has had to

• 1) establish Network Infrastructure between NIHs
  Research Infrastructure sites Biotechnology
  Resources and General Clinical Research Center
  sites. The BIRN Coordinating Center (BIRN CC)
  defined the infrastructure required and
  coordinated the procurement of both hardware and
  software installed at the BIRN Partner Sites
  (BIRN PS) for networking and caching of data.
• 2) establish policies and protocols for use of
  the BIRN infrastructure and operating guidelines.
  The BIRN CC is coordinating with BIRN PS groups
  and NIH - National Center For Research
  Resources.
• 3) build a Virtual Data Grid (VDG) for the
  initial cohort of group projects or test-beds
  in the neuroimaging research domain.
• 4) develop the data modeling and mediation tools
  for complex queries relevant to the proposed
  research goals of these test-bed projects.

11
BIRN Coordinating Center Activities / Layers

• I. Networking Infrastructure for the BIRN -
  Install, commission and tune networks between the
  BIRN Partner Sites with I2/Abilene etc.
• II. The BIRN Virtual Data Grid - Install and
  commission hardware and software layers for
  BIRN's Distributed Data - The Storage Resource
  Broker Technologies (SRB).
• III. Coupling BIRN Data Grid to the Computational
  Grid - Make the BIRN system work with NSFs
  National Middleware Initiative and leverage on
  the NSFs Distributed Terascale Facilities.
• IV. Data Integration Technologies for BIRN -
  BIRN-CC Tool Development and BIRN-Partner Site
  Group-Specific Data Respresentations and Domain
  Map Development
• V. Interaction Environments for the BIRN - Adding
  a Collaborative layer to the BIRN data complex
  query Portal, including Volumetric Viewing
  Capabilities and link to the BIRN-Partner Site
  visualization and analysis tools.

12
Networking and Data Storage Infrastructure
Site Hardware Configurations

• Last link included
• Linking Internet 2 to Biotechnology Resource or
  General Clinical Research Center sites is of
  primary concern
• BIRN supplies major networking pieces in BIRN
  Racks
• Site infrastructure to link from rack to Internet
  2 provided by each campus
• Each site has a consistent configuration
• A standard base for all network and data services
  to run on -
• No excuses - local staff to become responsible
  for their site
• Remote management and measurement tools at each
  site
• To facilitate expansion of BIRN, alternative
  lower cost configurations will be explored and
  specified - once test-bed projects are established

13
BIRN Network Operations Center (NOC)
Logical Description with Standardized Site Rack
24x7 Operations Monitoring
Internet 2
SDSC Network
BIRN NOC Monitoring

• Network Switch
• BIRN Statistics
• BIRN Network Probe
• Network Attached Storage
• Grid POP/ SRB, Globus Dual Processor
• General purpose ( eg., encryption)
• UPS for Rack

NCRR
Cisco 4006
GB Switches
DL380 - Network Stats
BIRN Network Management
GigE Net Probe
N2400 NAS 0.5- 4 TB
Gigabit Network Probes
DL380 Grid POP
DL380
NOC Servers and Testers/Analyzers

APC UPS
BIRN Site Rack
14
Building a Distributed Data Collection

• Provide data handling system to access data
  stored in someone elses administration domain
• Provide collection management to support
  discovery of distributed data sets
• Provide knowledge management interface to
  differentiate between naming conventions

15
BIRN Virtual Data Grid

• Defines a Distributed Data Handling System
• Integrates Storage Resources in the BIRN network
• Integrates Access to Data, to Computational and
  Visualization Resources

• Acts as a Virtual Platform for
  Knowledge-based Data
  Integration Activities
• Provides a Uniform
  Interface to Users

STORAGE RESOURCES
16
BIRN is Integrating Across Data Sources

• The idea
• Different models capture correlated but distinct
  aspects of biological reality
• How can we express and evaluate queries that
  compute data across models
• The approach
• For each source create a knowledge-base of the
  anatomy of the observations
• Attach the data of each source to the respective
  knowledge-base
• bridge the sources by a simple ontological
  mapping
• Compute the query across bridged sources

17
Data Sources on the Web
Washington Univ
Boston Univ
UCLA
Montana State Univ
Vanderbilt Univ
Yale Univ
18
Information Integration
What is the cerebellar distribution of rat
proteins with more than 70 homology with human
NCS-1?
Integrated View
Integrated View Definition
Mediator
Wrapper
Wrapper
Wrapper
Wrapper
Web
protein localization
morphometry
neurotransmission
CaBP, Expasy
19
Querying Across Databases
What is the distribution of protein X in
cerebellum?
Source 2
Source 1
Source 3
Cerebellum

• Purkinje Cell Dendrite
• Dendritic spines
• Dendritic shaft
• Endoplasmic reticulum

• Cerebellar Cortex
• Granule Cell Layer
• Purkinje Cell layer
• Molecular Layer

20
Additional Glue Knowledge
Source 2
Source 1
Source 3
Cerebellum
Cerebellar Cortex Granule Cell Layer Purkinje
Cell layer Molecular Layer
Purkinje Cell Dendrite Dendritic spines Dendritic
shaft Endoplasmic reticulum
21
(No Transcript)
22
Development of Smart Atlases
23
From Telescience to BIRN
24
Brain Morphometry BIRN

• Clinical Specific Aims
• Attempt to address questions of the following
  type
• Do structural differences contribute to specific
  symptoms such as memory dysfunction or depression
  independent of diagnosis?
• Do specific structural differences distinguish
  specific diagnostic categories?
• Technological Specific Aims
• Attempt to overcome obstacles to the use of
  neuroimaging data as quantitative outcome
  measures for clinical investigation including the
  issues raised by longitudinal and multi-site
  studies. In other words- to make the Clinical
  Specific Aims possible to address.

Harvard
Duke
UCLA
UCSD
25
Cortical Thickness Estimation with Sub-Voxel
Accuracy
Gray-white boundary
Pial surface
From Anders Dale / Mass General Hospital -
Harvard
26
A Surface-Based Coordinate System
byAnders Dale (MGH)
Subject (native representation)
Inflation (native representation)
Registration (common representation)
Return to Native Configuration (with coordinates)
27
Automated Whole-Brain Segmentation (37 Anatomical
Labels, 1 wk 10 min)
Part of Free Surfer
By Bruce Fishl and Anders Dale (MGH)
28
Brain BIRN Data Flow
29
4 x 10-5 mm3 110,000 X
5 mm3
30
The Mouse Brain Imaging Research Network MBIRN
Caltech/Beckman Inst.
31
BIRN test-bed sites, with 10 site project for
fMRIs of Schizophrenics - Multiscale Mouse
Models of Disease, Human Brain Morphometrics and
FIRST BIRN