A microsatellite is a short, noncoding DNA sequence (a Tandemly Repetitive DNA sequence) that is repeated many times within the genome of an organism. Many repeats tend to be concatenated at the same locus.
In a microsatellite, the repeated sequence is very simple, consisting of from two to four nucleotides, and can be repeated 10 to 100 times. Hence if one sequenced a microsatellite locus, one would expect to see a sequence of "TAGTAGTAGTAGTAGTAGTAG..."
The number of repeats at a particular locus is hypervariable (highly polymorphic) between individuals of the same species. It is for this reason that microsatellite sequences can be used for genetic fingerprinting and paternity testing. Most loci of the genome, even non-coding parts, would be too similar to allow individuals to be reliably distinguished.
The hypervariability arises because the repeated simple sequences cause a high frequency of loss or insertion of additional repeats by confusing the DNA replication machinery. Self-complementary sequences may aid this process.
In tumour cells, where controls on replication may be damaged, microsatellites may be gained or lost at an especially high frequency during each round of mitosis. Hence a tumour cell line might show a different genetic fingerprint from that of the host tissue.
See also: minisatellite, mobile element, transposon, short interspersed repetitive element, long interspersed repetitive element, junk DNA, selfish DNA