How to perform a microarray experiment

While musing at the thought of using microarrays, many questions pop up in one's mind: What type of arrays do I want to use? How do I design my experiments? Which labeling method do I want to use? How do I analyze my microarray data?

To obtain relevant results and perform a microarray experiment that best suits your needs, you should carefully set up a plan before starting at the bench. You will find help and inspiration in the various chapters. Regarding the design of DNA microarray experiments, we would like to refer to a review by Yang and Speed (7).

DNA microarrays can be classified according to the type of probes on the array (cDNA, oligonucleotides, genomic fragments), their generation and immobilization. In many cases presynthesized molecules (PCR products, oligonucleotides, isolated DNA) are deposited on the array either by contact printing (using metal pins that carry small volumes of probe solution due to capillary action) or by non-contact printing, when probe solution is dispensed by ink-jet printing. In addition, several companies generate high-density microarrays by synthesizing oligonucleotides in situ (for review see 8). The synthesis is either based on specific base deprotec-tion by light (coordinated by photomasks or digital micromirror devices) or on chemical deprotection and the use of ink-jet technology. An alternative to these commercial systems is the open-source platform POSaM (piezoelectric oligonucleotide synthesizer and microarrayer), described by Lausted et al. (9). These authors present the low-cost production of in situ synthesized oligonucleotide arrays (containing 9800 features) in their lab. An overview of commercially available oligonucleotide arrays is given in Table 1.1.

cDNA arrays can also be purchased (see Table 1.2), or produced as described in various chapters of this book (Chapters 2, 3, 4, and 5). In addition to protocols for the generation of genomic DNA arrays (see Chapters 11 and 12), you can find commercial suppliers in Table 1.2.

After the hybridization of labeled target molecules to DNA arrays, fluorescent signals are detected by laser scanners or CCD cameras. Chapters 15 and 16 describe image acquisition and image data conversion.

Finally, microarray data produced in large scale need to be stored and processed to obtain relevant and meaningful results. At this point the field

Table 1.1. Commercially available oligonucleotide microarrays

Principle of DNA microarray generation In situ synthesis

Photodeprotection using photomasks (~25mers)

Photodeprotection using digital mirrors (DMD) (24mers-70mers)

Chemical deprotection using ink-jet technology (60mers)

Presynthesized oligonucleotides spotted onto arrays

50mers on expoxy surface glass slides 80mers on coated glass slides Long oligomers

30mers on 3D-matrix-coated slides 30mers on 3D-matrix-coated slides 50mers on 3-micron beads 70mers on proprietary slide substrate



( NimbleGen#*


Agilent Technologies* (


BD Biosciences (Clontech) (

TeleChem International, Inc. (

Mergen Ltd. (

GE Healthcare (Amersham Biosciences) (

Illumina, Inc.


Microarrays Inc* (

Arrays for ChIP-on-chip assays (#) or array CGH (*) are also provided.

Table 1.2. Companies producing cDNA and genomic microarrays cDNA arrays

Miltenyi Biotec GmbH (

Scienion AG (

Takara Bio, Inc. (

Cambrex Bio Science (

Genomic arrays

Aviva Systems Biology (



Spectral Genomics, Inc (

Vysis (Abbott Laboratories) (

of bioinformatics found a vast playground and the increased use of microarrays triggered the co-evolution of various bioinformatics methods (Figure 1.2). In general, all data need to be normalized. Not to get you lost at this early step, several normalization methods are presented in Chapter 17. The special case of normalizing array CGH data is described in Chapter 21.

After normalization, further data processing depends on the questions you intend to address. If you want to compare gene expression under different biological conditions (e.g. normal vs. tumor tissue, wild type vs. knockout/transgene cells, control vs. drug-treated cells etc.) the first and foremost question concerns differential gene expression. Chapter 18 takes you on a safe trip to a list of differentially expressed genes.

DNA microarray time course experiments are highly suitable to monitor the expression of a very large number of genes during a biological process over a defined period of time (one example is given in Chapter 5). Chapter 20 describes the statistical analysis of time-course data.

Chapter 19 deals with clustering and classification. Clustering is, for example, used to find a group of genes, which have similar expression patterns or a group of samples (e.g. tissue samples from patients), which show likewise expression of a set of genes. Classification methods can determine whether a gene expression profile of a tissue sample belongs to a certain class, and are applied to predict disease courses.

Finally, special applications require special bioinformatic analyses. Therefore, the processing of data generated by single nucleotide polymorphisms (SNP) detection and ChIP-on-chip experiments is addressed separately in Chapters 8 and 13, respectively.

Detection of Single Nucleotide Changes

Array CGH


Gene Expression Profiling

Data Preprocessing / Normalization

Normalization Strategies 17

8 Contains a Section on Data Analysis in SNP Detection

Determining Differential Target Hybridization 18

Array CGH Data Analysis

11 Contains a Section on Array CGH Data Analysis

13 Contains a

Section on ChIP-on-chip Data Analysis

Clustering / Classification

Time Course Analysis

Figure 1.2.

Flowchart of bioinformatics methods that are used at different steps of microarray data analysis. Numbers indicate book chapters.

Ten years ago, DNA microarrays became fashionable because they enable high-throughput gene expression analyses. Over the years, they gained importance with their expanding use in medical diagnostics and research and even now scientists are continuing to advance this technology and its applications. The examples of ChlP-on-chip and array painting show that the combination of two techniques can lead to a new technology - so it will be exciting to see what's yet to come

0 0

Post a comment