Satellite DNA refers to long arrays of noncoding tandem repeat sequences that are abundant in eukaryotic genomes. Satellite DNA arrays are most concentrated in the centromere regions of chromosomes, but can also be interspersed throughout the chromosome or near the telomeres [1]. The term satellite DNA originated from genomic DNA density gradient centrifugation experiments, which revealed “satellite” bands of a different densities than the rest of the genome, which were found to be tandemly repeated sequences [2], [3]. The different density of satellite bands indicated biased nucleotide composition, which still holds true for satellite sequences discovered with modern techniques, with most being AT-rich (e.g.,[4], [5]). Among the first taxa studied with density gradient centrifugation were fruit fly species such as Drosophila melanogaster and Drosophila virilis, whose genomes were estimated to be made up of 10 % and 40 % satellite DNA, respectively [6], [7], [8]. Fruit flies have since been a leading model in studying satellite DNA biology. Although some authors consider microsatellites a type of satellite DNA, we will make a clear distinction in the definition we use here. Microsatellites have both short unit lengths (1–6 bp) and short array lengths (typically <50 bp in Drosophila). Satellite DNAs could have any unit length (in Drosophila they are typically <20 bp [4], [5]), but must be present in long arrays (typically at least kilobases long). Microsatellites have historically been more tractable to study variation in since they have clear boundaries and individual loci can be amplified and their length studied (e.g. [9]). Our review will focus on processes affecting satellite DNA in long arrays.

Ever since they were discovered in the genome, the role (or lack thereof) for satellite DNA has been questioned and debated [10]. Satellite DNA repeats do not code for proteins. However, biologists hypothesized they might have various biological roles, such as in regulating recombination, chromosome segregation, gene expression, structural roles, or roles in embryonic development that were not yet obvious [11], [12], [13], [14]. Another pervasive view was that satellite DNA represented junk DNA and behaved mostly neutrally with respect to the host [15], [16]. Indeed, satellite DNA sequences and abundances are not conserved across taxa like typical functional loci of the genome [17], which has supported the idea that they are likely nonfunctional to the host. There are several possible explanations for satellite DNA persistence and rapid evolution that do not require host-level function. Satellite DNA variants also have the potential to act selfishly, by biasing their inheritance into the egg rather than polar body during asymmetric female meiosis, known as centromere drive [18]. In this way, selfish centromeric satellite DNA arrangements or abundances could outcompete other haplotypes, even if they are not more favorable to the host (e.g.,[19]). Centromere drive can also explain the rapid evolution of satellite DNA sequences and abundances [20]. Furthermore, non-beneficial satellites may have a higher chance of persisting in regions near the centromere, since these regions have lower recombination rates, lowering the efficacy of (negative) selection [21].

Recent studies have started to reveal potential functions of satellite DNA, raising the possibility that changes in satellite DNA sequences/abundance can disrupt their cellular functions and activities. In this context, the fact that satellite DNA is rapidly evolving on evolutionary timescale has a couple of fundamental implications: (1) how can the underlying instability mechanisms explain rapid divergence of satellite DNA sequence/abundance during evolution?; (2) changes in satellite DNA may disrupt cellular processes involving them, which may be implicated in speciation and incompatibilities between closely related species. In this review, we will discuss satellite DNA instability in the context of copy number changes, and the processes that maintain satellite DNA but can also explain rapid divergence in an evolutionary scale. We will mostly discuss findings and implications that are relevant to Drosophila. Satellite DNA instability can also involve large-scale chromosomal changes, which have been discussed in other reviews [22].