DNA Microarray: Principle, Procedure, Types, and How to Read the Results

A single glass slide the size of a thumbnail can carry more than twenty thousand tiny DNA spots, one for almost every gene in the human genome. Wash a labeled sample across it, and each spot lights up in proportion to how active its gene was in that sample. In one experiment, you read the behavior of the entire genome at once. That is the idea behind the DNA microarray: instead of asking about one gene at a time, you ask about all of them together, and let each gene report itself by finding its matching partner on the chip.

What a DNA microarray is

A DNA microarray (also called a DNA chip or gene chip) is a slide or chip carrying thousands of microscopic spots arranged in a grid. Each spot holds many copies of a known, single-stranded DNA sequence called a probe, and each probe corresponds to a specific gene or DNA sequence. When labeled DNA from a sample is applied to the chip, it binds only to the spots whose probes it matches, a process called hybridization. Measuring how much sample bound at each spot tells you how much of that sequence was present in the sample.

The microarray is the chip and the assay. The instrument that reads it, exciting the fluorescent labels and measuring the light from each spot, is a separate device, the microarray scanner, covered in detail in our article on the microarray scanner. This article focuses on the microarray itself: how it works, its types, and how to interpret the result.

The principle: hybridization

The entire technique rests on one property of DNA: complementary base pairing. A single strand of DNA will bind specifically to another single strand carrying the complementary sequence, A with T, G with C, forming a stable double strand held by hydrogen bonds. This specific binding is called nucleic acid hybridization.

On a microarray, the probes are single-stranded and anchored to the chip. When single-stranded sample DNA (or DNA made from the sample's RNA) is washed over the chip, each sample fragment binds only to the spot whose probe is its complement. A fragment with no matching probe washes away. Because each spot's sequence is known in advance, the position of a bound, glowing spot tells you which gene it represents, and the brightness tells you how much bound.

The two-color comparison: reading the spots

Most classic expression microarrays are run as a two-color comparison, and understanding this design is the key to reading the result.

Setting up the comparison. Two samples are compared on one chip, for example, healthy tissue versus tumor tissue, or untreated versus treated cells. RNA is extracted from each, converted to complementary DNA (cDNA), and each sample's cDNA is labeled with a different fluorescent dye: one sample with Cy3 (which fluoresces green) and the other with Cy5 (which fluoresces red). The two labeled samples are mixed and hybridized together to the same chip, so at every spot the two samples compete to bind their shared probe.

Reading the spots. After washing off unbound cDNA, the scanner measures green and red fluorescence at each spot, and the color of the spot tells you which sample expressed that gene more:

A red spot means the gene was more active in the Cy5-labeled sample. More red-labeled cDNA bound, so that sample had more of this transcript.

A green spot means the gene was more active in the Cy3-labeled sample.

A yellow spot means both samples expressed the gene about equally. Roughly equal red and green combine to yellow.

A black or dark spot means neither sample expressed the gene; almost nothing bound.

The numerical version of this is the red-to-green intensity ratio at each spot, which quantifies how much more active a gene was in one sample than the other. Reading a microarray, then, is reading a field of colored spots as a genome-wide comparison of two conditions.

Procedure

The workflow from sample to data has four stages.

1. Chip preparation. The probes are attached to the chip in an ordered grid. Two main manufacturing approaches are used. In situ synthesis builds the probes base by base directly on the chip surface using photolithography and light-directed chemistry (the approach used for high-density oligonucleotide arrays). Spotted arrays instead take pre-made probes and deposit them onto the chip with a robotic printer. Many chips are now bought commercially ready-made.

2. Sample preparation and labeling. RNA is extracted from each sample, then converted to more stable complementary DNA (cDNA). During this step a fluorescent dye is incorporated, Cy3 or Cy5 for a two-color experiment. (Single-color platforms label one sample per chip and compare across chips instead.)

3. Hybridization. The labeled sample is applied to the chip and incubated, allowing fragments to find and bind their complementary probes. The chip is then washed to remove unbound and loosely bound material, leaving only specifically hybridized sample.

4. Scanning and analysis. The microarray scanner excites the dyes, measures the fluorescence at each spot, and produces a digital image. Software converts each spot to an intensity value, subtracts local background, and (in two-color experiments) computes the red-to-green ratio, turning the image into a table of expression values gene by gene.

Types of microarray

Although gene-expression arrays are the most familiar, the same hybridization principle is used for several distinct purposes.

Expression microarrays measure how actively genes are transcribed, by hybridizing labeled cDNA made from a sample's RNA. This is the two-color comparison described above, used to compare gene activity between conditions such as diseased and healthy tissue.

Genotyping microarrays (SNP arrays) detect single-nucleotide polymorphisms, single-base differences in the genome, rather than expression. The probes are designed so that a sample's DNA binds differently depending on which base it carries at a given position, allowing thousands to millions of genetic variants to be typed at once. These are used in genome-wide association studies and in clinical genetic testing.

Comparative genomic hybridization (CGH) arrays detect gains and losses of DNA (copy-number changes) rather than expression or single bases. Sample and reference DNA are labeled with two dyes and co-hybridized, and the ratio at each spot reveals whether a region is duplicated or deleted in the sample. Array CGH is widely used to find chromosomal imbalances in cancer and in constitutional disorders.

The common thread across all three: known probes on a chip, labeled sample DNA, and hybridization read out as fluorescence. What changes is what the probes are designed to detect.

Uses of DNA microarray

Microarrays are used wherever many sequences must be measured in parallel. In gene-expression profiling, they compare which genes are active between conditions, for example, identifying genes switched on in a tumor versus normal tissue. In cancer classification, expression patterns help distinguish tumor subtypes that look similar under the microscope but behave differently, informing prognosis and treatment choice. In pharmacogenomics and drug research, they reveal how cells respond to a drug at the level of gene activity. In genetic testing and research, SNP and CGH arrays detect variants and copy-number changes linked to disease. In microbiology, arrays can be designed to detect and identify pathogens or resistance genes by their sequences.

Limitations of DNA microarray

It measures what you put on the chip. A microarray can only detect sequences for which probes were designed. Genes or variants not represented by a probe are invisible to it. This is a key difference from RNA sequencing, which can detect transcripts without prior knowledge of their sequence.

Relative, not absolute, quantification. A microarray measures fluorescence intensity, which is quantitative, but it usually reports relative expression, how much more active a gene is in one sample than another, rather than an absolute transcript count. Results are often confirmed with quantitative PCR.

It measures transcripts, not proteins. Expression arrays report mRNA levels. They cannot capture what happens after transcription, so a gene that is transcribed but not translated into protein, or a protein regulated after it is made, will not be reflected. Microarray data describes transcription, not the final protein output.

Reproducibility and cross-hybridization. Signal can vary between runs, and closely related sequences can bind the wrong probe (cross-hybridization), adding noise. Careful normalization and validation are needed.

How to Remember

The chip binds, the scanner reads. The microarray does the biology, probes catching complementary sample DNA by hybridization. The scanner does the physics, exciting dyes and measuring light. Keep the two jobs apart and the topic gets simpler.

Red, green, yellow: who won the spot. Two samples compete at each spot. Red means the red-labeled sample expressed that gene more, green means the green-labeled sample did, yellow means a tie, black means neither. The color is a scoreboard for each gene. To keep the dyes straight: Cy3 is green, Cy5 is red, so the lower number is the cooler color.

Hybridization is just A-T, G-C, at scale. The whole technique is complementary base pairing repeated across thousands of spots. Every spot is one small hybridization experiment, and the chip runs them all at once.

Probes are chosen in advance, so you only see what you look for. A microarray answers questions you designed probes for. Anything without a probe is invisible. That one fact explains its biggest limitation and how it differs from sequencing.

Key exam facts in one table

Where Students Get Confused

The microarray is not the scanner. The microarray is the chip and the assay. The scanner is the instrument that reads it. They are different things, and this article is about the chip.

Which color means what. In a two-color experiment, a red spot means the gene was more active in the sample labeled with the red dye (Cy5), and green means more active in the Cy3 sample. Yellow is equal expression, not a separate gene state. Students often forget the color points back to which sample, not to the gene itself.

Hybridization happens before scanning. By the time a chip is scanned, sample has already bound to probes and the chip has been washed. Scanning only measures the result; it does not cause the binding.

Probes are known; samples are unknown. The sequence at each spot is designed in advance. What you are measuring is how much of the unknown sample binds there. The chip is the reference; the sample is the question.

Microarray versus RNA-seq. A microarray only detects sequences it has probes for. RNA sequencing reads whatever transcripts are present without needing probes designed first. This is why sequencing has largely overtaken microarrays for discovery, though arrays remain useful for targeted, standardized testing.

Expression, SNP, and CGH arrays are not the same. All use hybridization, but they detect different things: gene activity, single-base variants, and copy-number changes, respectively. The word microarray alone does not tell you which.

References

1. Brown TA. Genomes 5. 5th ed. Boca Raton: CRC Press; 2023.
2. Buckingham L. Molecular Diagnostics: Fundamentals, Methods, and Clinical Applications. 4th ed. Philadelphia: F.A. Davis; 2026.
3. Wiltgen M, Tilz G. DNA microarray analysis: principles and clinical impact. Hematology. 2007;12(4):271-287. doi:10.1080/10245330701283967
4. National Human Genome Research Institute. DNA Microarray Technology Fact Sheet. genome.gov; 2020. Available at: https://www.genome.gov/about-genomics/fact-sheets/DNA-Microarray-Technology