T.

Mohapatra
NRC on Plant Biotechnology
IARI, New Delhi
DNA Microarray
A microarray is like a massive RNA or
protein blotting experiment...in reverse
Labeled
RNA or protein
TARGET(s)
Fixed surface
- microscope slide
- silicon chip
Short DNA or
protein (e.g. antibody)
PROBES
Basic principles
Main novelty is one of scale
hundreds or thousands of probes rather
than tens
Probes are attached to solid supports
Robotics are used extensively
Informatics is a central component at all
stages
Major technologies
cDNA probes (> 200 nt), usually
produced by PCR, attached to either
nylon or glass supports
Oligonucleotides (25-80 nt) attached to
glass support
Oligonucleotides (25-30 nt) synthesized
in situ on silica wafers (Affymetrix)
cDNA chips
Probes are cDNA fragments, usually amplified
by PCR
Probes are deposited on a solid support,
either positively charged nylon or glass slide
(poly-L-lysine coated microscope slides)
Attachment is commonly through ionic
bonding
Samples (normally poly(A)+ RNA) are
labelled using fluorescent dyes
At least two samples are hybridized to chip
Fluorescence at different wavelengths measured
by a scanner
Technology requires no bioinformatic
preselection but weak attachment chemistry
Yields increased background following
hybridizations
Probes are not gene-specific
cDNA chips
cDNA chip design
Probe selection
Non-redundant set of probes
Includes genes of interest to project
Corresponds to physically available clones
Chip layout
Grouping of probes by function
Correspondence between wells in
microtitre plates and spots on the chip
Probe selection
Make sure that database entries are cDNA
Preference for RefSeq entries
Criteria for non-redundancy
>98% identity over >100 nt
Accession number is unique
Mapping of sequence to clone
Use Unigene clusters
Directly use data from sequence verified collection
(e.g. Research Genetics)
Independently verify sequence
cDNA arrays on nylon and glass
Nylon arrays
Up to about 1000 probes per filter
Use radiolabeled cDNA target
Can use phosphorimager or X-ray film
Glass arrays
Up to about 40000 probes per slide, or 10000 per
2cm
2
area (limited by arrayers capabilities)
Use fluorescent targets
Require specialized scanner

Long oligonucleotide probes are designed based upon
fine-scale bioinformatic analyses of genome or EST
databases
Probes are usually targeted to 3 UTRs and are gene-
specific
Oligos are synthesized with 5- or 3-amine modification
that binds covalently to e.g. aminosilane-coated slides
Cross-hyb amongst family members is minimized
Attachment chemistry forms strong bonds
Hybridization T
m
is uniform and background is relatively
low
Long oligo (70mer) Arrays

Glass chip manufacturing
Choice of coupling method
Physical (charge), non-specific chemical, specific
chemical (modified PCR primer)
Choice of printing method
Mechanical pins: flat tip, split tip, pin & ring
Piezoelectric deposition (ink-jet)
Robot design
Precision of movement in 3 axes
Speed and throughput
Number of pins, numbers of spots per pin load
Spotted Array Technology
Labeling and hybridization
Targets are normally prepared by oligo(dT)
primed cDNA synthesis
Probes should contain 3 end of mRNA
Need CoT1 DNA as competitor
Specific activity will limit sensitivity of assay
Alternative protocol is to make ds cDNA
containing bacterial promoter, then cRNA
Can work with smaller amount of RNA
Less quantitative
Hybridization usually under coverslips
Standard protocol for
comparative hybridization
Single-channel Array Hybridization
Two-channel Array Hybridization
Scanning the arrays
Laser scanners
Excellent spatial resolution
Good sensitivity, but can bleach fluorochromes
Still rather slow
CCD scanners
Spatial resolution can be a problem
Sensitivity easily adjustable (exposure time)
Faster and cheaper than lasers
In all cases, raw data are images showing
fluorescence on surface of chip
Zeptosens : Planar Waveguide Principle - for High
Sensitivity Fluorescence Microarray Detection
free label
Imaging
of surface-confined fluorescence
excitation of bound label
CCD camera
microarray
on chip
Image Capture
Function is to convert digital information into a viewable
image linked to quantitative data in each channel for
each feature (cDNA or oligo). Will subtract local
background and compute ratiometric data.
Affy system uses GeneChip Operating Software (GCOS)
For spotted array systems, many many software options!
Most popular is arguably the GenePix Pro image analysis
software.
The Affymetrix approach
Probes are oligos synthesized in situ using a
photolithographic approach
There are at least 5 oligos per cDNA, plus an
equal number of negative controls
The apparatus requires a fluidics station for
hybridization and a special scanner
Only a single fluorochrome is used per
hybridization
It is very expensive !
Affymetrix chip production
NimbleGen Arrays
Maskless Array Synthesizer (MAS) Technology
Commercial chips
Clontech, Incyte, Research Genetics - filter-
based arrays with up to about 8000 clones
Incyte / Synteni - 10000 probe chips, not
distributed (have to send them target RNA)
Affymetrix - oligo-based chips with 12000
genes of known function (16 oligos/gene)
and 4x10000 genes from ESTs
Alternative technologies
Synthesis of probes on microbeads
Hybridization in solution
Identification of beads by fluorescent bar
coding by embedding transponders
Readout using micro-flow cells or optic
fiber arrays
Production of universal arrays
Array uses a unique combination of oligos,
and probes containing the proper
complements
Arrays for genetic analysis
Mutation detection
Oligos (Affymetrix type) representing all
known alleles
PCR followed by primer extension, with
detection of alleles by MALDI-TOF mass
spectroscopy (Sequenom)
Gene loss and amplification
Measure gene dosage in genomic DNA by
hybridization to genomic probes
Bioinformatics of microarrays
Array design: choice of sequences to be used
as probes
Analysis of scanned images
Spot detection, normalization, quantitation
Primary analysis of hybridization data
Basic statistics, reproducibility, data scattering,
etc.
Comparison of multiple samples
Clustering, classification
Sample tracking and databasing of results
Microarrays A Statisticians Feast
Chip layout design (mainly for spotted arrays)
how many replicates of each probe?
where to spot or print on the array?
Non-biological variability introduced through
labeling and hybridization steps
Signal variability per feature
amongst replicate probes within chips (spotted
arrays)
dyes within chips (two-channel array experiments)
amongst chips (one-channel and two-channel)
Data analyses via multivariate statistical
approaches
Experimental Design
Two-channel (spotted array) experiment 2 targets
e.g. steady-state wild type versus mutant
wild type
x
Chip 2
Chip 1
chip replicate
wild type
x
mutant mutant
wild type
x
mutant
wild type
x
wild type
x
mutant
dye swap
Chip 3
Chip 4
chip replicate
Experimental Design
Two-channel (spotted array) experiment 3 targets
THE LOOP DESIGN

dye swap
wild type
x
mutant-1
wild type
x
mutant-1
mutant-1
wild type
x
dye swap
wild type
x
mutant-1
mutant-1
x
mutant-2
mutant-1
x
mutant-2
mutant-2
mutant-1
x
mutant-1
x
mutant-2
wild type
x
mutant-2
wild type
x
mutant-2
mutant-2
wild type
x
wild type
x
mutant-2
dye swap
Normalization Issues
Unequal RNA/cDNA used
Different labeling concentrations
Hybridization techniques leading to different rates of
bonding
Gene expression data have meaning only in the
context of the particular biological sample and the
exact conditions under which the samples were taken.
For instance, if we are interested in finding out how
different cell types react to treatments with various
chemical compounds, we must record unambiguous
information about the cell types and compounds used
in the experiments
Normalization Techniques
Use housekeeping genes as standards
too much fluctuation in biological
systems
Median of all signal intensities good
approximation
Combination of all samples most
accurate
ROOT (7-days-old seedlings) ROOT (7-days-old seedlings)
Root-specific genes
Seed-specific genes
R
O
O
T

(
7
-
d
a
y
s
-
o
l
d

s
e
e
d
l
i
n
g
s
)

S
E
E
D

(
0
-
2
D
A
P
)

A B
ROOT (7-days-old seedlings)
S
E
E
D

(
0
-
2

D
A
P
)

Genes showing low level but
highly differential expression
Identification of Seed-specific Genes Using Microarray
(100)
(50)
(10)
(2)
Analysis of the Normalized Intensity Data
Many, many software options for analyses of microarray data.
Employ an application that uses SOUND STATISTICS, especially one
capable of ANOVA.
ANOVA is the only way that you can evaluate statistical significance of
expression data by comparing more than 2 observations simultaneously
(i.e. goes beyond simple ratios).
ANOVA captures variability introduced by dye bias (spotted arrays) and
chip-to-chip differences amongst replicates: The more variability, the
higher the p-value on the expression ratio, and the less you should trust
the result.
Good applications that employ ANOVA:
Bioconductor (freeware based in R programming language)
SAS Microarray Suite (commercial offering in SAS language)
S+ArrayAnalyzer (commercial offering in S programming
language)
Uses of Microarrays
Genome-scale gene expression analysis
Differentiation
Responses to environmental factors
Disease processes
Effects of drugs
Detection of sequence variation
Genetic typing
Detection of somatic mutations (e.g. in
oncogenes)
Direct sequencing
Uses of Microarrays
Transcriptome profiling:
Massive parallel analysis can reveal patterns of gene expression
allowing researchers to predict gene associations and pathways
Microarrays have been used over the past five years for
transcriptome analyses in organisms ranging from bacteria to
fungi to humans to higher plants under hundreds of different
conditions
Diagnostics: RNA or DNA samples hybridized to small arrays to
detect expression of marker genes (stress/pathogens,
developmental events)
Molecular genetics: Genomic DNA samples are hybridized to a
limited number of probes to detect single nucleotide polymorphisms
(SNPs) for mutational analyses (pedigrees, gene mapping)
Gene promoter analyses: ChIP to Chip assays overlapping or
tiled DNA probes representing an entire genome are hybridized
with labeled, putative trans-acting proteins to detect binding to cis
elements in promoter regions
Microarray data on the Web
Many groups have made their raw data
available, but in many formats
Some groups have created searchable
databases
There are several initiatives to create
unified databases
EBI: ArrayExpress
NCBI: Gene Expression Omnibus
Companies are beginning to sell microarray
expression data (e.g. Incyte)
Web links
Leming Shis Gene-Chips.com page very
rich source of basic information and
commercial and academic links
DNA chips for dummies animation
A step by step description of a microarray
experiment by Jeremy Buhler
The Big Leagues: Pat Brown and NHGRI
microarray projets