Image Quantitation in Microarray Analysis

1
Image Quantitationin Microarray Analysis
Moretomorrow...
2
Microarray analysis

• Array construction, hybridisation, scanning
• Quantitation of fluorescence signals
• Data visualisation
• Meta-analysis (clustering)
• More visualisation

3
Technical
image from Jeremy Buhler
4
Experimental design

• Track whats on the chip
• which spot corresponds to which gene
• Duplicate experimental spots
• reproducibility
• Controls
• DNAs spotted on glass
• positive probe (induced or repressed)
• negative probe (bacterial genes on human chip)
• oligos on glass or synthesised on chip
  (Affymetrix)
• point mutants (hybridisation plus/minus)

5
Images from scanner

• Resolution
• standard 10?m currently, max 5?m
• 100?m spot on chip 10 pixels in diameter
• Image format
• TIFF (tagged image file format)
• can be compressed
• (eg. Lempel-Ziv-Welch 5x compression)
• 1cm x 1cm image at 16 bit 2Mb (uncompressed)
• other formats exist eg. SCN (used at Stanford
  University)
• Separate image for each fluorescent sample
• channel 1, channel 2, etc.

6
Images in analysis software

• Typical experiment
• normal state, Cy3-labelled sample (green)
• perturbed state, Cy5-labelled sample (red)
• hybridisation, then scanning
• overlay images ? pseudo-colour image
• qualitative representation of results
• Image spot colour Signal strength Gene
  expression
• yellow normal perturbed unchanged
• green normal gt perturbed repressed
• red normal lt perturbed induced

7
Quantitation process (1)

• Accurate representation of signal for each
  spotand determine ratio channel1channel2
• Determine spot boundary
• construct grid (dimensions of array / spot size)
• iterative process to find spots
• Measure signal
• fluorescence
• 8 bit 256 shades
• 16 bit 65536 shades
• absolute output values vary from system to system

8
Quantitation process (2)

• Measure background
• local (usually best)
• selected region
• selected spots / probes from different species
• Quality control
• eg. fraction of pixels greater than background
  (ScanAlyze)
• flag aberrant spots
• Determine ratio of signal strengths for each
  spot Ch1/Ch2 (Ch1I-Ch1B)/Ch2I-Ch2B)

9
Normalisation

• Eliminate systematic variation
• correct for
• dye incorporation
• print-tip effects
• hybridisation efficiencies
• etc.
• How?
• Use

STATISTICS!
10
Normalisation

• Simple example green vs red
• Normgreeni greeni

TotRed - TotRedB TotGreen - TotGreenB
11
Normalisation

• More complexe
• Within slide
• global (constant over the slide)
• spot intensity dependent (spot by spot)
• within print-tip group (group by group)
• scale (outlayers spread)
• Between slides
• paired-slides (dye swap)
• Multiple slides
• scale (slide outlayers spread)

12
Normalisation
13
Normalisation

• Example Within slide print-tip group scaling

• Before scaling

• After scaling

14
Normalisation references
statistics

• Normalization for cDNAmicroarray data
• Yang et al. (2001)
• In preparation
• Statistical methods for identifyingdifferentially
  expressed genes in replicatedcDNA microarray
  experiments
• Dudoit et al. (2000)
• Technical report 578 Berkeley Statistics Dept.
• Both PDFs are available from the web site of the
  course

15
Quantitation - problems

• Reference signal is close zero
• channels ratio (Ch1/Ch2) tends to infinity
• Non-uniform background
• mean background sometimes non-representative
• bright particles
• streaks on image
• safer to use median (middle value)
• less contribution by extreme values

16
Background problems
17
Background removal
18
Quantitation à la ScanAlyze
signal
background
19
ScanAlyze output

• CH1I ch1 intensity
• CH2I ch2 intensity
• SPIX number of pixels in spot
• CH1B median intensity of the local background
  (recommended)
• CH2B median intensity of the local background
  (recommended)
• CH1BA mean intensity of the local background
• CH2BA mean intensity of the local background
• BGPIX number of background pixels
• Thus to calculate channel ratios Ch1 CH1I -
  CH1B
• --- ----------- Ch2 CH2I - CH2B
• Quality control
• CH1GTB1 fraction of pixels in spot greater than
  background (CH1B)
• CH2GTB1 fraction of pixels in spot greater than
  background (CH2B)
• CH1GTB2 fraction of pixels in spot greater than
  1.5 X background (CH1B)
• CH2GTB2 fraction of pixels in spot greater than
  1.5 X background (CH2B)
• CH1EDGEA mean magnitude of the horizontal and
  vertical Sobel edge vectors within spot 1
• CH2EDGEA mean magnitude of the horizontal and
  vertical Sobel edge vectors within spot 2

20
(No Transcript)
21
Input files for Cluster

• Table tab delimited text, 1 line/gene, 1
  column/experiment

• Minimal table

• Extended table

22
Prepare data for Cluster

• Exp1

• Exp2

. . .

• ExpN

23
Software packages - quantitation

• ScanAlyze
• by Michael Eisen (Stanford University)
• quantitation of images
• no data visualisation
• free from http//rana.lbl.gov/
• ImaGene
• BioDiscovery Inc.
• quantitation and some data visualisation
• demo from http//www.biodiscovery.com/
• plus many others - explore!

24
Making sense of raw data

• Difficult to see results in tabulated data
• Represent in graphical form
• Data visualisation examples from
  ImaGene and others. . .

25
Data visualisation - scatter plot
26
Data visualisation - M vs A
27
Data visualisation - pie chart
28

• ScanAlyze
• quick demo