- 20 Apr 2007

Working Page to standardize some of the MAGE entries across Mouse BIRN or BIRN test beds

1. the identifier prefix: To identify where the data is from and what source (Accession number)
   • is there a BIRN standard we should adopt? It is in 2004 to construct MAGE object identifiers in the following LSID-like manner: <authority>:[<namespace>]:<object>[:<revision>>]
   • Recommendation: Jeff G?
2. database reference code: This specifies if the raw data is coming from another DB or if we point to data in another DB.
   • Recommendation: Use the DB reference codes specified by EBI research http://www.ebi.ac.uk/~ele/ext/submitter.html#db (#5)
3. acceptable data file format
   • Currently BIRN Microarray DB accepts CSV format, people may expect Excel, XML, or some other format for PROCESSED data (after normalization)
   • File format may depend on if we decide to accept data earlier in the workflow, i.e. a .cel file (reference Daniel Sforza's presentation)
   • Recommendation: .CSV, .XML (later), .CEL(Affymetrix CEL file)
4. log entries information
   • After upload, the provenance information should be stored, possible examples: IP address, error messages,
   • Recommendations: We need input on what other types of information should be kept in this area

Ontology Entries

Click herefor entries need to be mapped to BIRNLex.

1. Information about who is providing the data and where they are from
   • Role (Submitter/Experimenter)
   • Primary Investigator
   • Institute name, url
   • Funder and Grant #
   • Contact Information (phone, email)
   • Other?
2. File information
   • File name
   • File location on other DB (url, uri, or urn)
   • File format
   • Date file generated (need this?)
   • Other?
3. Experiment Information
   • Experiment name
   • Experiment description
   • Array Chip Type
   • Annotation or Reference to an external annotation database
   • Normalization or Summarization
   • Public accessible?
   • Publication Status (published, unpublished, submitted, in press, confidential)
4. Experiment details
   • IRB information or human subject information
   • Group of Interest (normal, control, diseased, pre-diseased)
   • Treatment methods
   • Treatment chemical/compound
   • Tissue or structures information
5. Subject(s) information
   • Species (preferred name and scientific name)
   • Strains
   • Developmental Stage (embryonic stages and adult stages)
   • Age/Age Unit
   • Sex(male, female, hemphrodites, other)

Ontology references:

• BIRNLEX Ontology
• MGED Ontology, mapping to MAGE-OM Model
• Mouse Anatomical Dictionary
• OBO Foundry Ontologies

---

---

MAGE-TAB

MAGE-TAB is a simplied version of MAGE-ML, and is proposed to be part of MAGEv2. The data is arranged in tabular format instead of XML like MAGE-ML. It consists four different types of files: Investigation Description Format(IDF), Array Design Format(ADF), Sample and Data Relationship Format(SDRF), Raw and processed data files.

• Click MAGE Group or A simple spreadsheet-based, MIAME-supportive format for microarray data: MAGE-TAB for MAGE-TAB v1.0 specification and further info.
• Compare MAGE-TAB and Tab2MAGE by clicking here
• Tools
  • convert MAGE-TAB to MAGE-ML
  • convert MAGE-ML to MAGE-TAB

Examples and Use Cases

Investigation Description Format (IDF) Example:

Investigation Title	Chronic Mouse Model of Parkinson's Disease
Experimental Design	Parkinson_Disease_Design
Experimental Factor Name	genetic_modification_design
Experimental Factor Type	genetic modification
Experimental Facotr Term Source REF	BIRNLEX
Person Last Name	Sforza
Person First Name	Daniel
Person Email	sforza@ucla.edu
Person Phone	xxx-xxx-xxxx
Person Address
Person Affiliation	Laboratory of Neuro Imaging
Person Roles	Experimenter
Person Roles Term Source REF
Quality Control Type
Quality Control Term Source REF
Replicate Type
Replicate Term Source REF
Date of Experiment	yyyy-mm-dd
Public Release Date	yyyy-mm-dd
PubMedID
Publication Author List
Publication Status
Experment Description
Protocol Name
Protocal Type
Protocol Description
Protocol Parameters
Protocol Term Source REF
SDRF File	e-external_file.csv
Term Source Name	BIRNLEX
Term Source File	http://fireball.drexelmed.edu/birnlex/
Term Source Version	1.2

Sample and Data Relationship Format(SDRF) Example

Sourc Name	Hybridization Name	Array Design REF	TERM SOURCE REF	Array Data File	Derived Array Data File
Sample 1	Experiment 1	A-AFFY-MOE430	AFFYMETRIX_TERM	Data1.CEL	AFFY_data1.CHP

Sample Name	Characteristics [Organism]	Characteristics [Organism Part]	Characteristics [Organism Part]	Characteristics [Developmental Stage]	Protocol REF	Parameter Value [Compound]	Extract Name	Labeled Extract Name	Label	Hybridization Name
M6_FCX	Mouse	Brain	Frontal Cortex	Adult		MPTP	M6_FCX_Ext	M6_FCX_lab_biotin	Biotin	M6_CTx_hyb
C6_FCX	Mouse	Brain	Frontal Cortex	Adult		Saline	C6_FCX_Ext	C6_FCX_lab_biotin	Biotin	C6_CTX_hyb

---

---

Tab2MAGE

This is a software package supported by ArrayExpress curators, which is able to generate MAGE-ML from Tab2MAGE spreadsheet document.The spreadsheet consists of three sections, separated by one or more blank lines: Experiment, Protocol and Hybridization.

• Examples
• Click here for Term Definition
• Tools
  • convert Tab2MAGE to MAGE-TAB
  • convert Tab2MAGE to MAGE-ML

---

---

GEO SOFTmatrix and GEOarchive

Daniel has submitted a microarray dataset to GEO using the SOFTmatrix and GEOarchive metadata format, and found it much clearer with example and annotations for large Affymetrix datasets. Their format may be easier for researchers to organize their information easily, and easily integrate with the database entries.

Example

• Click here for Daniel's submittied file, which also includes the mapping between GEOarchive meta data and MAGE-TAB metadata.
• Link to GEO site
• Metadata mapping between GEOarchive and MAGE-TAB.

  GEOarchive	MAGE-TAB
  SERIES	IDF
  title	Investigation Title
  summary	Experiment Description
  overall design	Experimental Design
  type	N/A
  contributor[]	Person Last Name Person First Name Person Mid Initial
  PLATFORM
  title	Investigation Title
  technology	Protocol Contact
  distribution	N/A
  organism	Characteristics[] in SDRF section
  manufacturer	Term Source
  manufacture protocol	Term Source File
  SAMPLES	SDRF
  Sample name	Sample Name
  title	Hybridization Name
  CEL file [Affymetrix submissions only]	Array Data File
  EXP file [Affymetrix submissions only]	Array Data File
  source name	Source Name, Material Type
  organism	Characteristics[organism]
  characteristics[]	Characteristics[] (e.g Characteristics[Strain], Characteristics[Gender], Characteristics[Age])
  molecule	Extract Name, Material Type
  label	Label
  description	Description
  platform	reference to Term Source Name in IDF file
  PROTOCOLS
  growth protocol	Protocol Name, Protocol Type, Protocol Description
  treatment protocol	Investigation Title
  extract protocol	Same as Above
  EXP file [Affymetrix submissions only]	Same as Above
  label protocol	Same as Above
  hyb protocol	Same as Above
  scan protocol	Same as Above
  data processing	Normalization Name
  value definition	N/A

---

---

MAGE-ML

Example XML broken down by MAGE-ML package

Audit and Security

Audit and Security package contains classes that describe an individaul contact information, or an oragnization information, as wells as the the security group they belong, and the role they play.

Classes	Item	Detail	Example
Audit and Security
Contact	Organization	name
		affiliation
		roles (ontology term)
	Person	name (last, first)
		affiliation
		roles (ontology term)
		phone
		email
		fax
Property Set	IRB	number
		exp date
		grant number
	Public rec?	yes or no

Description

Description package contains classes that describe the external databases or annotations being referenced to.

Classes	Item	Detail	Example
Annotation and Description
	Databases
	Database Entries
	Ontology Entries

Array

Array package contains classes that describe individual arrays, including detailed information on relevant manufacturing proccesses. It also contains references to the LIMS data that might contain BioMaterial information being used.

Classes	Item	Detail	Example
Array
ArrayManufacture	ID	unique identifier for each manufacturer, use manufacturer assession number as prefix
	Person	(Same as Contact->Person)
	Arrays	contain multiple array
Array	ID	unique identifier for each array chip
		other array properties should be automatically loaded based on the array chip model

ArrayDesign

ArrayDesign package contains classes that describe a microarray design. PhysicalArrayDesign describes the design that is being used to manufacture physical array.

Classes	Item	Detail	Remarks and Example
Array Design
Physical Array Design	Composite Group
	PhysicalArrayDesign
	DesignElemntGroup	FeatureGroup	Contain multiple Feature
		Species	ontology term

DesignElement

DesignElement contains classes that describes the purpose of each Array. The Feature describe the intended location on the Array. It can be specified as reporters or CompositeSequence for the arrays.

Classes	Item	Detail	Example
Design Element
Feature	Controlled Features
	Control Type

Experiment

Experiment package contains a collection of BioAssays that are related to the ExperimentalDesign. ExperimentalDesign is the description and collection of ExperimentalFactors and the BioAssays information.

Classes	Item	Detail	Example
Experiment
Experimental Factors
	Factors
	Providers
ExperimentDesign

BioSequence

BioSequence is a representation of DNA, RNA, protein sequence. It can be represented as Clone, Gene or a sequence.

Classes	Item	Detail	Example
BioSequence

BioMaterial

BioMaterial package describes the biological material being used, and the description of the creation through BioSource, BioSample, LabelExtract classes. BioMaterial can be related to other BioMaterial through DAG(directed acyclic graph). BioSample are products of treatments that are of interest. It can be used as the sources of other BioSamples. BioSource is the raw material information. LabelExtracts are special BioSamples^? that have Compounds that are detectable.

Classes	Item	Detail	Example
BioMaterial
	BioSource
	BioSample
	Other BioMaterial
	ActionMeasurement
	LabelExtract

BioAssay

BioAssay represents both physical and computational groupings of arrays and biomaterials. PhysicalBioAssay is a bioAssay created by the BioAssayCreation event. A measured bioAssay is the direct processing of information in a physical bioAssay by the featureExtraction event. DerivedBioAssay is created by the Transformation BioEvent^? from one or more MeasuredBioAssays or DerivedBioAssays.

Classes	Item	Detail	Example
BioAssay	DerivedBioAssay
	MeasuredBioAssay
	PhysicalBioAssay
	BioAssayCreation
	Hybridization
	BioAssayTreatment
	FeatureExtraction
	Image
	Image Acquisition
	Channel

BioAssayData

BioAssayData describes gene expression data. BioDataValues contain the actual values.

Classes	Item	Detail	Example
BioAssayData
	DesignElements
	BioAssays
	QuantitationTypes

Protocol

Classes	Item	Detail	Example
Protocol

QuantitationType

QuantitationType describes the data that are communicated in microarray experiments. It can be Standard or Specialized types. The standard types can be MeasuredSignal, DerivedSignal, Ratio or PresentAbsent, or PValue, Error and ExpectedValue. Specialized types are user defined.

Classes	Item	Detail	Example
QuantitationType

---

---

Case Reference

• Daniel Sforza: Microarray data Workflow Presentation
• Rob William: Illumina33Tissues Tab2MAGE format

---

Working Page to look at the Minimum Amount of Metadata a Microarray Data Submitter should provide

---