DNA microarray
/ MY-kro-array /
Suppose you want to ask thousands of questions at once: is gene 1 present or active? gene 2? gene 3? ...all the way to twenty thousand. A DNA microarray is a way to ask all those questions simultaneously on a single glass slide, by laying out thousands of known DNA probes in a tidy grid and seeing which ones a sample sticks to.
The chip is a grid of tiny spots; each spot is densely packed with many copies of a known single-stranded DNA probe, and each spot's identity (which gene it represents) is fixed by its location. You take your sample — say, all the RNA from a cell, converted to fluorescently labeled cDNA — and wash it over the chip. Wherever a sample molecule finds a spot with a matching sequence, it base-pairs and sticks (this is hybridization). After washing away the unbound material, you scan the chip: the brightness of each spot reports how much matching sequence was in the sample. Read the whole grid and you have, in one shot, a readout across thousands of genes.
Microarrays were the first technology to measure the activity of a whole genome's worth of genes at once, and they powered a wave of expression profiling, genotyping, and comparing healthy versus diseased tissue. Their key limitation, and the reason RNA-seq has largely overtaken them, is that an array can only detect what its probes were designed for — you cannot discover something unexpected, and the relationship between brightness and amount is not perfectly quantitative. They are 'closed' questionnaires of fixed questions, whereas sequencing is an open-ended interview.
On a gene-expression array, the spot for an immune gene glows brightly in an infected sample but stays dim in a healthy one — instantly showing that gene is switched on during infection, alongside the simultaneous readout of thousands of other spots.
Each grid spot is a known probe; brightness reports how much matched.
A microarray can only see what its probes were designed to detect — it cannot discover the unexpected, and its signal is semi-quantitative, which is why open-ended RNA-seq has largely replaced it for gene-expression studies.