Skip to content.


2008 AHM (Oct.20) MBAT3.0 Requirements

This document lists the required functionality for MBAT that will be delivered at the 2008 All-hands Meeting on October 20,2008. For each category, the collaborators are also listed.


Search Core

  1. Design client data source plugin framework [Daren,Queenie,Steve]
    1. will plugins work with current mbSearch core?
  2. Implement client data source plugin framework [Queenie]
  3. Migrate each datasource as plugin [Queenie]
    1. evaluate what info on metadata server is needed
  4. Query Results Table: [Steve]
    1. Add move/drag column (already exists)
    2. Add sort by column
    3. Add customizable columns (ie: right click table headers and select columns to display like Windows Explorer)
    4. Add resizable JPanel for table
    5. Thumbnail view

Do these concurrently or wait till framework is finished?

  1. Complete support for all query terms
  2. Add CCDB datatypes (wait for plugins?)
    1. 3D images (returned as analyze images?)
    2. 3D reconstructed surfaces?

Do we have time for this?

  1. Add keyword search functionality
    1. identify which datasources can have keyword functionality
    2. how does term source API integrate?
    3. what will be result types and how to display in results table?
    4. design GUI (need clustering in the future)
  2. Add RDF query language functionality
    1. SPARQL (CCDB Cellular Knowledge Base: http://purl.oclc.org/ckb/)

Viewer Core

  1. CCDB Zoomify
    1. annotation functionality
  2. Smart Atlas
    1. stereotaxic coordinates

Analysis Core

  1. Spatial Registration Plugins
    1. Landmark Based Warping (Davit)
    2. Jibber (Asif)

DocManager Core

  1. ImageJ File Plugins

Open Issues

  1. ImageJ processing plugins
  2. Atlas interoperability
  3. Plugins Embedded?
    1. allow plugin to create JFrames vs. everything embedded in MBAT?
  4. Capture work
    1. XML save file

Collaborative Efforts

  1. CCDB/AIDB
  2. Universal WebServices API


Task List

  1. Reevaluate AHM Requirements
    1. Determine new features to add by Fall AHM
      1. Keyword search
      2. Analysis component in viewer
      3. Framework for clients to add new Data Sources
      4. Access control module for client sources
      5. Public vs private databases
      6. Landmark Registration
    2. Determine collaborative efforts to support
      1. Uniform API (XCEDE, FUGE/MAGE, WebServices)
      2. AIDB / 3D Data sources
      3. Concept Mapper
      4. NeuroTerrain
  2. Redesign entire MBAT framework
    1. Code review for mbSearch core
    2. Merge old and new mbSearch datatypes (ie: ProbeNew?, etc)
    3. Refactor all icons to common/resources/icons. Check build scripts for unused or duplicate (nested) jar includes.
    4. Shiva.IO classes to standalone module
    5. Data Source data in MetaData or XML



Existing MBAT Improvements

  1. Automatically sync to appropriate atlas slice for given slice
    • Also show a marker for location of smaller relative size datasets in atlas

Data Sources [Caltech, Duke, UCSD, CC, UCI, fBIRN, XCEDE WG]

  1. Add 3D images (AIDB)
    1. Determine webservices API [Ilya,Jeff B.,Allan, Jeff G, XCEDE WG]
    2. Determine data exchange format (XCEDE) [need a subgroup from Mouse to work more with XCEDE WG]
  2. Fix ROI Query (ArcIMS? or new image server? Asif works with access, but not sure who are directly responsible for servers)
    1. Server-side problems?
    2. Needs to be easier to find data
    3. Needs more data in database
    4. Needs more metadata about data retrieved

Atlasing [UCSD, Drexel]

  1. Infrastructure for Atlas Interoperability Server
  2. Importation of Drexel's NeuroTerrain Atlas into MBAT format
    • next steps are to use this framework to access MBL data using their atlas and Atlas Interoperability infrastructure

Query Engine [AIDB, all Mouse DBs, Brian, Vadim, and Bill]

  1. Refactor
    1. Clean up v2.0
    2. Gene expression to mbSearch
      1. MBAT GUI code must change
  2. Issues
    1. How to classify Brain mosaics in query terms
  3. Query Terms
    1. Experimental Info
      1. Experimental Type
        1. Any
        2. Gene Expression
          1. Any
          2. Image based
          3. Microarray
        3. MRI
        4. Electron Microscopy
        5. Cell Fills
      2. Project
        1. Any
        2. BxD
        3. Development
        4. Reeler
        5. EAE
        6. Sex Differences/Dimorphism
        7. Keyword... (freetext)
      3. Investigator
        1. Any
        2. Keyword... (freetext)
    2. Gene/Molecule
      1. Gene Attribute
        1. Gene Symbol (freetext)
        2. Gene Name (freetext)
        3. Protein (freetext)
        4. Keyword... (freetext)
      2. rather than the more specific gene function and we'll have to connect to a more powerful annotation DB-both Rob's and Paul Pavlidis'(?)
      3. Add way to list available gene symbols?
    3. Subject/Organism
      1. Species
        1. Mouse
      2. Strains
        1. C57BL
        2. 12956
        3. A/J
        4. B6D2F1
        5. BALB/c
        6. CBA
        7. C3H
        8. DBA
        9. FVB
      3. Developmental Stage
        1. default: adult
        2. All/any
        3. Embryonic (0-11 embryonic days)
        4. Fetal (11-20(birth) embryonic days)
        5. Neonate/infant (0-12 postnatal days)
        6. Child/pup (12-24 postnatal days)
        7. Juvenile (24-60 postnatal days)
        8. Adult (60-300 postnatal days)
        9. days/weeks/months?
      4. Gender
        1. any
        2. male
        3. female
        4. other
    4. Anatomical Area
      1. Any
      2. CNS hierarchy
      3. Entire Organism
      4. Cell
      5. Subcellular
      6. Other part (freetext)
  4. Query Term Dependencies
    1. If Gene Expression selected --> show Gene/Molecule searchgroup
    2. If Gene/Molecule search group enabled --> grey MRI
    3. If MRI selected --> disable Gene/Molecule searchgroup
  5. Formulating query
    1. Integrate with Concept Term Mapping
    2. Integrate with Mediator
  6. Expand current data sources
    1. AIDB
      1. Query terms?
      2. Steps required for MBAT to connect to AIDB?

Registration Workspace [UCSD, Drexel]

  1. Framework for loading data
    1. from Query workspace
    2. locally
  2. Framework
    1. All registration "external" to MBAT
      1. Infrastructure for local binaries (Java & non-Java)
      2. Infrastructure for remote binaries (Pipeline, UCSD Portal)
    2. Infrastructure for data types:
      1. subimage
      2. 2D
      3. 3D
      4. 4D
    3. Infrastructure for landmark based algorithms
    4. Infrastructure for intensity based algorithms? this may be where we can create a tie to Pipeline
  3. Supported Data types
    1. 2D - 2D
      1. manual
      2. default algo: Ryan and Craig's algorithm, taking advantage of Heng's UI for points (possibly contours)
    2. Subimage - 2D
      1. default algo: manual
    3. Subimage - 3D
      1. default algo: manual
    4. 2D - 3D
      1. default algo: ???: _facilitate manual first-find close spot in atlas, then possible implementation of Erh-Fang's contour-based algorithms, although it may be too difficult to implement easily, then follow up with 2D - 2D using algorithm and interface developed above_

Comparison Workspace

  1. Framework for loading data
    1. locally
    2. from Registration workspace
    3. pre-registered spec (XML)
    4. from SRB? or is this covered by Registration workspace
  2. Framework for flexibility and customization
    1. Allow for N comparisons of datasets
    2. Allow for customizable window layout and size
    3. Allow for virtual page of infinite size
  3. Framework for side-by-side view
  4. Framework for maximizing real estate for comparison image
    1. Replace 4-up view
      1. Side-by-side: two windows
      2. Overlay: one window if this is an option that's fine, but not a default as we've found that in general people prefer side-by-side
    2. Collapsible subwindows
      1. Thumbnail viewer
      2. Comparison cart?
    3. Context driven controls
      1. HUD controls overlayed over image for zoom/pan/rotate (aka: GoogleEarth?)
  5. Analysis Tools
    1. Linked controls
      1. Handle different scales? (ie: 2 volumes with different sizes)
      2. Handle different data types? (ie: 2D v. 3D image / surface v. image)
    2. Viewing
      1. View XY, YZ, XZ planes
      2. Pan
      3. Annotate
      4. Browse through slices
      5. Zoom
    3. Overlay comparison ???
      1. Blending (Remove background color / need mask?)
      2. Logical operations (union, intersection, etc)
      3. Layers (aka: GoogleEarth?)