Standards and Data Handling with Microarrays

1
Standards and Data Handling with Microarrays

• David J Craigon

2
What is he going to talk about?

• Data handling- some database stuff
• followed by
• Standards- the MGED group of standards

3
What is he going to talk about?

• Data handling- some database stuff
• followed by
• Mostly standards- the MGED group of standards
• MIAME
• Ontology
• MAGE

4
Some Databases

• ArrayExpress- the EMBL of Microarrays- MAGE based
• MIAMEExpress
• Various species specific databases

5
The Microarray Gene Expression Data Society (MGED)

• A standards body for microarrays
• Split into 4 working groups
• MIAME
• Ontology
• MAGE
• Data processing

http//www.mged.org
6
Why have standards?

• Standards allow us to
• Be prescriptive (Oi you, do this)
• Aid communication between people
• Produce best practices
• Enables replication of experiments

7
The Microarray Gene Expression Data Society (MGED)

• A standards body for microarrays
• Split into 4 working groups
• MIAME
• Ontology
• MAGE
• Data processing

http//www.mged.org
8
The MIAME Standard

• Pronounced like the city. Miami.
• An abbreviation for Minimum Information about A
  Microarray Experiment

9
MIAME

• Scientific papers should contain all information
  required to replicate an experiment
• What information do you need for a microarray
  experiment?
• People talk of documents being MIAME-compliant.

10
A guide to MIAME
11
World of Microarrays

• Array Design
• Experimental Design
• Samples used, extract preparation and labelling
• Hybridization procedures and parameters
• Measurement data and specifications of data
  processing

12
Array Design
13
ArrayArray as a whole
Platform Type
Dimensions of the array
Surface and coating specifications
Number of Features
Availability of array, or production protocol
14
Some terminology

• Any given slide/genechip/whatever is an Array
• The design of the array (which spot goes where,
  and where all of the spots are) is an Array
  Design
• Any spot etc. on the array is a Feature
• What the feature is made of is a Reporter

15
Features

• A feature is a specific spot on an ArrayDesign
• Record
• Feature dimensions
• How the feature is attached to the array
• Where each individual feature is on the
  ArrayDesign

16
Reporters

• A reporter is a specific probe
• You could spot a reporter multiple times on one
  array, or even on multiple arrays

17
Reporters

• Record
• What type of reporter it is
• Single or double stranded
• The sequence if available
• If not PCR primers, approximate length
• How this reporter was made
• What it reports for (database reference)

18
Special features

• Control features
• How is it a control

• Composite features
• How combined?
• What does it report for?

19
Samples used, extract preparation and labelling
20
Biosource and Biosample

• Biosources and Biosamples are either
• The original, living things that the RNA was
  extracted from
• Or an extract from that organism
• e.g. Patient, Mouse, mouse brain, RNA extract are
  all Biosamples/Biosources.
• Biosources are the original material you start
  with- biosamples are all subsequent steps.

21
Biosourcerecord
Contact details for sample (where did you get it
from?)
Mouse (?)
Arabidopsis
Human
Organism
22
More BioSource

• Relevant descriptors such as
• Sex
• Age
• Developmental stage
• Organism part
• Cell type
• Plant strain or line
• Genetic variation
• Individual genetic characteristics
• Disease state
• Additional clinical information
• Which individual?

23
Progression of a microarray experiment
Protocol
Protocol
Protocol
Protocol
Seed
Plant
Cold plant
RNA extract
Labelled RNA extract
BioSource
BioSample
BioSample
BioSample
LabelledExtract
24
Protocols

• Between each step is a protocol
• e.g.
• Growth protocol
• Sample treatment protocol
• Separation technique
• Extraction protocol
• Labelling protocol

25
Extraction protocol

• Extraction method
• RNA, mRNA, genomic DNA extracted
• Amplification steps

26
Labelling Protocol

• Amount labelled
• Label used
• Label incorporation method
• Spikes type, qualifier (e.g. concentration,
  expected ratio), the elements on arrays that
  match the spike

27
Hybridisation Protocol

• Which samples were hybridised to which arrays
• Solution used
• Blocking agent
• Wash procedure
• Quantity of labelled extract used
• Time, concentration, volume, temperature
• Description of hybridisation instruments

28
Measurement Data and Specifications of Data
Processing

• Three steps
• Scan
• Raw analysis
• Summary analysis

29
Scanning protocol

• Scanning Hardware, Software used
• Scan Parameters used e.g. PMT voltage
• Scanned images (?)

30
Image analysis protocol

• Image analysis software and version (e.g. GenePix
  version 4)
• Description of the algorithm used
• Parameters used
• All output from the analysis

31
Normalised and Summary data

• Data processing protocol
• Gene expression data tables derived from the
  experiment as a whole.
• Derived measurement value
• Reliability indicators

32
Experiment

• About the complete experiment

33
Experiment description

• Which Arrays are part of the experiment?
• Which Samples are part of the experiment?

34
Experiment Design

• Contact details for the experiment
• Type of the experiment (gene knock-out,
  treatment)
• Arrays and experimental factors
• Quality control steps taken
• Description for the experiment

35
No more MIAME!
36
The Microarray Gene Expression Data Society (MGED)

• A standards body for microarrays
• Split into 4 working groups
• MIAME
• Ontology
• MAGE
• Data processing

http//www.mged.org
37
Computer Standards

• Computers are a lot less forgiving than people
• Computer standards allow lots of different
  software to work together

38
Computer standards

• People are beginning to realise that it is useful
  for computers to understand databases, as well as
  people

39
Ensembl example
40
Advantages of data standards

• Use many databases together
• Better automated searching
• One piece of software can use many databases
  interchangeably

41
What is an ontology?

• An ontology is like a dictionary- it defines many
  terms
• An ontology is a restricted dictionary- it
  attempts to avoid duplication
• An ontology is a structured dictionary- it states
  how the terms relate to each other

42
Demonstration of Gene Ontology, MGED ontology
43
The Microarray Gene Expression Data Society (MGED)

• A standards body for microarrays
• Split into 4 working groups
• MIAME
• Ontology
• MAGE
• Data processing

http//www.mged.org
44
MAGE

• Computer standard
• A standard object model for Microarrays
• Describes the various parts of a microarray
  experiment in a standard way

45
MAGE-ML

• A collection of these objects in an XML format
• One file can represent an entire experiment, part
  of an experiment, many experiments, or part of
  many experiments
• Example- Affymetrix supply ArrayDesign objects in
  a file to describe all of their chips

46
The end!

• Ive shown you
• Some microarray databases you can get data from
• MIAME
• Ontologies
• MAGE