Exporting MAGEML from the Stanford Microarray Database Using Perl MAGEstk - PowerPoint PPT Presentation

1 / 9
About This Presentation
Title:

Exporting MAGEML from the Stanford Microarray Database Using Perl MAGEstk

Description:

Exporting MAGE-ML from the Stanford Microarray Database Using Perl MAGE-stk ... Generated by single method to ensure uniformity ... – PowerPoint PPT presentation

Number of Views:49
Avg rating:3.0/5.0
Slides: 10
Provided by: johnm161
Category:

less

Transcript and Presenter's Notes

Title: Exporting MAGEML from the Stanford Microarray Database Using Perl MAGEstk


1
Exporting MAGE-ML from the Stanford Microarray
Database Using Perl MAGE-stk
  • Jeremy Gollub and Tina Boussard
  • Department of Genetics
  • Stanford University
  • smd.stanford.edu

2
Decomposing the problem
  • Manageable chunks (Meta packages?)
  • Experiment Meta Data
  • Array Design Data
  • Actual Data
  • Mapping
  • Relationships were hard-coded
  • Close correspondences were easy (Person)
  • Bad matches were problematic (FactorValues)
  • Data fantasies were unavoidable (Protocols)

3
Problems with Software and Model
  • Software/Model Source and targets
  • Hybridization, for example
  • sourceBioMaterialMeasurements
  • physicalBioAssayTarget
  • Documentation How to tell the XMLWriter that
    your data files are external


  • my w BioMAGEXMLWriter-gtnew
  • (fh gt\STDOUT,
  • external_data gt1,
  • Tells writer data is external
  • cube_holds_path gt1)
  • Tells writer bioDataCube has the file path
  • Usage Referencing other outside files

4
Instantiating and Writing Objects
  • Abstract classes
  • MAGE-stk should warn
  • Independence/ownership
  • Not all identifiable objects are independent
  • Objects do not get written if their owner is not
    present

5
Instantiating and Writing Objects
  • Non-identifiable objects (e.g., Describable)
    cannot live on their own
  • ExperimentDesign needs an Experiment object to
    belong to.
  • The experiment Object has to be passed to the
    MAGE object, not the ExperimentDesign Object
  • Create a blank Experiment object, and add the
    ExperimentDesign, and pass the Experiment object
    to the MAGE object.

6
Where to start Work Flow
  • Paying attention to the intended use case
    actually helped!
  • Additional workflows and use cases would be
    appreciated.

7
General Approach
  • Perl MAGE-stk
  • Package and object-level exporter classes, as
    appropriate
  • Print ? ArrayManufacture PhysicalArrayDesign
  • User ? Person
  • OntologyEntry exporter for CV objects
  • SeqType ? biosequencetype OntologyEntry
  • Utility module to export static objects
    (SMDUtils.pm)
  • Channels
  • QuantitationTypes
  • Databases
  • Generic method for generating identifiers

8
Identifiers
  • Roughly adhere to LSID standard
  • Generated by single method to ensure uniformity
  • For machine parsing (name attribute is for human
    readers)
  • Namespace identifier
  • smd.stanford.edu for SMD objects
  • Other for Database objects, as appropriate
  • URNLSIDebi.ac.ukdatabaseEBI
  • Object identifier
  • Contains one or more SMD object attributes
  • May contain non-database human information
  • MAGE-OM object appended, to escape degeneracy
  • Examples (URNLSIDsmd.stanford.edu)
  • print.printid802/MAGE.PhysicalArrayDesign
  • spotlist.printid_spotlist.spot802_17/MAGE.Feature
  • print.printid_at.gene model802_At3g9247/MAGE.Comp
    ositeSequence

9
Wisdom, Painfully Gained
  • Memory usage is high remove limits.
  • Performance is low report progress and dont
    run over the web! (XMLWriter could usefully
    become verbose)
  • Ontologies and best practices need some community
    effort (but for now, were doing what EBI tells
    us to do).
Write a Comment
User Comments (0)
About PowerShow.com