• Other zoomified 2D datasets
  • Alan's MBF Virtual Slice Data
  • CCDB/NCMIR large scale brain mosaics
  • MBL high-resolution scanned image

CONCEPT

Query Initiation:

Return to Query:

• • System used by NeuroCommons to host SPARQL query service

• Use SPARQL query using Jena API

Workflow (as of 04/07):

1. Middle layer object searching NC written by Queenie
2. This design was evaluated in the context of the general concept based queries (discussion under 4/6/07) and it seems it should fit nicely with this approach

Objective: To integrate the NeuroCommons Google Map API ABA demo into MBAT and if successful, use this same method for other types of data

Query Initiation:

1. Provide a user with a context-sensitive menu (i.e. right-click on Windows; CNTL-click on Mac) over a gene "painted" image overlaying microarray data on our atlases.
   • That menu enables a user to select a related 2D data set based on metadata run against the ABA RDF repository such as:
     • Gene Symbol or name
     • Slice orientation if both are available (sagittal or coronal)
     • Brain region (once NeuroCommons has finished populating their map of which sections cut through which ABA atlas brain region volumes)
     • Additional categories such as project or imaging type for data from other sources
2. The conceptual global query interface may be used to generate a query in a similar manner as above

Return to Query:

• With the returned collection of ImagePyramid URLs, MBAT will present a Java browser using the NeuroCommons Google Map JS code
  • Initially defaulting to the median section in the series (coronal or sagittal)
  • Later want to connect this with our atlasing infrastructure
• Add to the current Google Map widgets a slice-axis slider to enable the user to select other section image pyramid sets ("stained" & "expressed").
  • This widget would simply enable the user to navigate across the list of section URLs (ImagePyramid URLs in Alan R's ABA ontology) for a given sectioned brain (coronal or sagittal) for a given gene.
  • We could also enter a "Slice Axis" toggle button (for those 4000 brains where ABA has both coronal & sagittal) to enable a user to flip between the two views.

General Plan:

• Obtain the ABA Zoomify image set URLs linked to a given gene name
  • System used by NeuroCommons to host RDF SPARQL query service
    • OpenLink Virtuoso Universal Server - RDF Database & SPARQL

• Use these within the Google Map AJAX/JS code in MBAT
• Use RDF SPARQL query formalism see Jena for SPARQL use
  • Jena RDF library
  • SPARQL use

Technical decisions include:

1. Determine which Java library provides the best Java HTML browser class
2. How to integrate instantiating that class in response to the user right-clicking on a gene name (or brain region)
   • including how to send the appropriate ABA zoomify image set URL into the JS code.
   • This particular task will be much easier once we have a Portal version of MBAT (WOMBAT), since then you'd simply have a Servlet/JSP page that contained the JS code and you could use various AJAX libs built for use with the Apache Tomcat Servlet container and Struts framework.
3. how to slip that window seamless into MBAT

link to NeuroCommons homepage

Steps taken:

4/07

1. SARPQL examples written by Bill
2. ABA architecture overview by Queenie includes diagram and tasks and required skills:
   • architecture diagram from Queenie:
     
   • Tasks and the required skills:
     • TWriting code to access ABA Server using SPARQL through web services from Virtuoso Server or through Jena ARQ API.
       • Use pre-defined SPARQL with the valid ontology terms to query from the repository.
       • Skill set: SPARQL (like SQL)
     • Writing java code to parse the SPARQL output data, and populate the image to the browser using Google MAP API.
       • Skill: Google MAP API and AJAX.
       • http://www.w3.org/TR/rdf-sparql-XMLres/
     • Integrate with MBAT
       • Skill: Java, and familiar with MBAT architecture.
3. Code for searching NC written by Queenie
4. This was design was evaluated in the context of the general concept based queries (discussion under 4/6/07) and it seems it should fit nicely with this approach