Aucun titre de diapositive
Download
Report
Transcript Aucun titre de diapositive
Introduction to
DNA Microarrays
DNA Microarrays and
DNA chips resources
on the web
INTRODUCTION
Microarray analysis is a new technology which
allows scientists to detect thousands of genes in a
small sample simultaneously and to analyze the
expression of those genes.
Microarrays are simply ordered sets of DNA
molecules of known sequence. Usually rectangular,
they can consist of a few hundred to hundreds of
thousands of sets. Each individual feature goes on
the array at precisely defined location on the
substrate.
Potential application domains
•
•
•
•
•
•
•
•
•
•
Identification of complex genetic diseases
Drug discovery and toxicology studies
Mutation/polymorphism detection (SNP’s)
Pathogen analysis
Differing expression of genes over time, between
tissues, and disease states
Preventive medicine
Ability to subtype disease and design drugs that treat
disease causes, rather than symptoms
Specific genotype (population) targeted drugs
More targeted drug treatments – AIDS
Genetic testing and privacy
The technique
Based on already known methods, such as fluorescence
and hybridization.
It's goal is to compare gene transcription in two or more
different kinds of cells.
There is four main steps in making an array experiment :
1- Array fabrication
2- Sample preparation and hybridization
3- Scanning the array
4- Exploring the results
The challenge
The big revolution here is in the "micro" term. New
slides will contain a survey of the human genome
on a 2 cm2 chip! The use of this large-scale method
tends to create phenomenal amounts of data, which
have then to be stored, processed and analyzed.
As the technique is quite new, analyzing the data is
still a problem, and nothing is standardized yet. A
few databases and on-line repositories are coming
out, and the future standard will probably be chosen
between these ones.
This is a job for…Bioinformatics !
Course overview
Theory :
Introduction to the technique of microarrays
Analyzing the data : a few methods
Quick survey of tools, databases, datasets
available on the web
Practical :
Using on-line tools for microarrays
Searching public databases
THE EXPERIMENT : making the chip
1- Designing the chip : choosing genes of
interest for the experiment
- Selection of chip probes that represent the
investigated genes.
- Finding sequences, usually in the EST
database.
- Problems : sequencing errors, alternative
splicing, chimeric sequences,
contamination…
THE EXPERIMENT : making the chip
2- Spotting the probes on the substrate
- Substrate : usually glass, but also nylon
membranes, plastic, ceramic…
- Probes : cDNA probes (500-5000 nucleotides,
dna chips), oligonucleotides (20~80-mer oligos,
oligo chips), genomic DNA ( ~50’000 bases)
- Printing methods : microspotting, ink-jetting
(for dna chips) or in-situ printing, for example
photolithography (for oligos, Affymetrix
method)
THE EXPERIMENT : making the chip
Microspotting and ink-jetting
THE EXPERIMENT : making the chip
The microspotting and ink-jetting are done by a
robot called “arrayer”
THE EXPERIMENT : making the chip
Oligo-spotting (Affymetrix method)
THE EXPERIMENT : hybridization
Sample preparation
- Extracting DNA (for genomic studies)
or mRNA (for gene expressions
studies) from the two samples to
compare.
-
Target labeling. Making cDNAs with
both extracts, and labeling them with
different fluors to allow direct
comparison.
THE EXPERIMENT : hybridization
Samples are eluted on the chip, put
in a hybridization chamber, and then
washed.
THE EXPERIMENT : generating data
Chip scanning
- Fluorescence measurements are made
with scanning laser fluorescence
microscope that illuminates each DNA
spot and measures fluorescence for
each dye separately. It creates one red
and one green image.
- The two images are then superimposed
to give a virtual result of RNA amounts in
both samples
THE EXPERIMENT : generating data
Chip scanning
1- Samples
2- Extracting mRNA
3- Labeling
4- Hybridizing
5- Scanning
6- Visualizing
Examples of scanner outputs
Affymetrix chip
Stanford chip
THE EXPERIMENT : generating data
Image analysis
-These fluorescence measures are then used
to determine the ratio, and in turn the relative
abundance, of the sequence of each specific
gene in the two mRNA or DNA samples.
- This analysis is performed by a software
such as “scanalyze”, available at :
http://rana.lbl.gov/EisenSoftware.htm
or “Spotfinder” from TIGR
- The files created can then be submitted to
further analysis
THE EXPERIMENT : making sense of the data
Although the visual image
of a microarray panel is
alluring, its information
content, per se, is still not
readable.
How can one visualize, organize and explore the
meaning of information consisting of several million
measurements of expression of thousands of genes
under thousands of conditions?
THE EXPERIMENT : making sense of the data
Data mining depends on the questions which
are asked. The most frequent question is to find
sets of genes that have correlated expression
profiles (belonging to the same biological
process and/or co-regulated), or to devide
conditions to groups with similar gene
expression profiles ( for example divide drugs
according to their effect on gene expression.
The method used to answer these questions is
called CLUSTERING.
Clustering data
•Input: N data points, Xi, i=1,2,…,N (the
color ratios measured with Scanalyze, for
example) in a D dimensional space. N and D
will be either genes and conditions for gene
clustering, or conditions and genes for
conditions clustering.
•Goal: Find “natural”groups or clusters.
•Note: according to the method, the number
of clusters will be fixed from the beginning
(K-means) or determined after the analysis
(hierarchical clustering)
Clustering data
Before clustering, a few steps to “clean the data
are necessary ( normalization, filtering)
Clustering methods :
1- Agglomerative Hierarchical
2- Centroids: K-means or SOM
3- Super-Paramagnetic Clustering
For a good introduction on different clustering techniques, read
the article from Gavin Sherlock “Analysis of large-scale gene
expression data” in Current Opinion in Immunology 2000,
12:201-205 (html,pdf)
Agglomerative Hierarchical Clustering
Distance between joined clusters
4
2
5
3
1
1
3
2
4
Dendrogram
5
The dendrogram induces a linear ordering
of the data points
Agglomerative Hierarchical Clustering
Before doing a hierarchical clustering, one has to
define two things :
1-The similarity measure between two genes (or experiments)
Centered correlation
Uncentered correlation
Absolute correlation
Euclidean
2- The distance measure between the new cluster and the others
Single Linkage:
distance between closest pair.
Complete Linkage: distance between farthest pair.
Average Linkage: distance between cluster centers
Centroid methods - K-means
•Start with random position of
K centroids.
•Iteratre until centroids are
stable
•Assign points to centroids
•Move centroids to center
of assign points
Iteration = 0
Centroid Methods - K-means
•Start with random position of
K centroids.
•Iteratre until centroids are
stable
•Assign points to centroids
•Move centroids to center
of assign points
Iteration = 1
Centroid Methods - K-means
•Start with random position of
K centroids.
•Iteratre until centroids are
stable
•Assign points to centroids
•Move centroids to center
of assign points
Iteration = 3
Self-organizing Maps
-Choose a number of partitions
- Assign a random reference
vector to each partition.
- Pick a gene randomly and
assign it to its most similar
reference vector.
- Adjust that reference vector is so
that it is more similar to the
chosen gene.
-Adjust the other reference
vectors.
- Repeat thousands of times until
partitions are stable.
A self-organizing map.
Super-Paramagnetic Clustering (SPC)
M.Blatt, S.Weisman and E.Domany (1996) Neural Computation
• The idea behind SPC is based on the physical
properties dilute magnets.
• Calculating correlation between magnet
orientations at different temperatures (T).
T=Low
Super-Paramagnetic Clustering (SPC)
M.Blatt, S.Weisman and E.Domany (1996) Neural Computation
• The idea behind SPC is based on the physical
properties dilute magnets.
• Calculating correlation between magnet
orientations at different temperatures (T).
T=High
Super-Paramagnetic Clustering (SPC)
M.Blatt, S.Weisman and E.Domany (1996) Neural Computation
• The algorithm simulates the magnets behavior
at a range of temperatures and calculates their
correlation
• The temperature (T) controls the resolution
T=Intermediate
Clustering data
Available clustering tools
•M. Eisen’s programs for clustering and
display of results (Cluster, TreeView)
–Predefined set of normalizations and
filtering
–Agglomerative, K-means, 1D SOM
•Matlab
–Agglomerative, public m-files.
•Dedicated software packages (SPC)
•Web sites: e.g.
http://ep.ebi.ac.uk/EP/EPCLUST/
•Statistical programs (SPSS, SAS, S-plus)
• And much others …
Clustering data
The final data representation
is then a big matrix with rows
being the genes and
columns representing the
different experiments. To
keep the image coherent
with the scan output, the
ratio numbers calculated by
Scanalyze are transformed
back in color spots on a
green-red based scale.
Clustering data
Another way to
represent these
data is a graph
showing the
gene’s expression
variation during
the different
experiments
Expression variation of nine genes along the
19 experiments from Lyer et al. (Fibroblast
response to serum stimulation)
Web resources : data analysis tools
Expression Profiler Online clustering and analysis tools
GenEx
Database, repository and analysis tools
MAExplorer
MicroArray Explorer for data mining Gene
Expression, free download
ArrayDB
Downloadable tools, short online demo
MAXD
Downloadable data warehouse and
visualisation for expression data
Jexpress
Java tools for gene expression data
analysis, free download
Michael Eisen's suite for image quantitation
and data analysis (Scanalyze, Cluster,
TreeView). Downloadable.
Eisen Lab
Web resources : public databases
SMD
The Stanford Microarray Database
Chip DB
Searchable database on gene expression
ExpressDB
Public queries of E. coli and yeast data
GEO
RAD
Gene expression data repository and online
resource
RNA Abundance Database
Expression
Connection
Saccharomyces Genome Database
expression data retrieval
EpoDB
Expression information retrieval for one
gene at a time
Public queries of yeast data
yMGV
Web resources : public databases
AMAD
Downloadable web driven database system
ArrayExpress
Public data deposition and public queries
maxdSQL
Downloadable data warehouse and
visualisation environment
GXD
Mouse expression data storage and
integration
Distribution and visualization of gene
expression data from any organism
GeNet
Web resources : public databases
Drosophila microarray
project
Drosophila Metamorphosis Time Course
Database
Samson Lab
Yeast Transcriptional Profiling
Experiments
NCBI SAGE data and analysis tools
SageMap
NCI60 cancer project
Serum-response
Breast cancer
Cancer Molecular
Pharmacology
Supplement to Ross et al. (Nat Genet.,
2000).
Supplement to Lyer et al.(1999) Science
283:83-87
Supplement to Perou et al. Nature
406:747-752(2000)
Integration of large databases on gene
expression and molecular
pharmacology.
Web resources : general information
Leung’s
Link’s page & softwares’ info
Davison’s
DNA Microarray Methodology - Flash
Animation
gene-chips
Overview of the technique, papers…
Chips &
microassays
General information
SMD guide
Stanford's links page, very complete
Introduction
Online introduction to microarrays
Microarrays protocols and arrayer
Brown Lab Guide
construction.