Minisatellites include some of the most variable loci in the human genome and are superb for dissecting processes of tandem repeat DNA instability. Single DNA molecule analysis has revealed different mutation processes operating in the soma and germline. Low‐level somatic instability results in simple intra‐allelic rearrangements. In contrast, high frequency germline instability involves complex gene conversions and is therefore recombinational in nature, almost certainly occurring at meiosis. To determine whether true meiotic crossovers occur at human minisatellites, we have used polymorphisms near the repeat array to recover recombinant DNA molecules directly from sperm DNA. Analysis of minisatellite MS32 has revealed an intense and highly localised meiotic crossover hotspot centred upstream of the array, the first example of a human hotspot defined at the molecular level. This hotspot extends into the beginning of the repeat array, resulting in unequal and equal crossovers. Array crossovers occur much less frequently than array conversions but appear to arise by a common process, most likely by alternative processing of a recombination initiation complex. The location of MS32 at the boundary of a recombination hotspot suggests that this locus has evolved as a by‐product of localised meiotic recombination activity, and that minisatellites might in general mark recombinationally proficient hotspots or hot domains in the genome. Finally, sperm crossover analysis makes it possible to explore the molecular rules that govern human meiotic recombination, and to detect phenomena such as meiotic drive that could provide a possible connection between recombination and DNA sequence diversity itself.

Minisatellites provide not only the basis for DNA fingerprinting and DNA profiling but also extremely informative systems for analysing processes of tandem repeat turnover in the human genome. Minisatellite instability appears to involve distinct mutation processes in somatic and germline cells; in the germline, mutation is frequently dominated by inter-allelic conversion-like events most likely occurring at meiosis and apparently regulated by cis-acting mutation initiator elements. Attempts to define these initiators in transgenic mice have so far been thwarted by what appears to be a major human/mouse barrier to the inter-species transfer of repeat instability. Minisatellites not only show high frequency spontaneous mutation in the germline, but also appear to be very sensitive to mutation induction by ionizing radiation, both in experimentally irradiated mice and in human populations exposed following the Chernobyl disaster; the mechanisms of mutation induction by radiation remain enigmatic.

Dispersed repeat elements contribute to genome instability by de novo insertion and unequal recombination between repeats. To study the dynamics of these processes, we have developed single DNA molecule approaches to detect de novo insertions at a single locus and Alu-mediated deletions at two different loci in human genomic DNA. Validation experiments showed these approaches could detect insertions and deletions at frequencies below 10 76 per cell. However, bulk analysis of germline (sperm) and somatic DNA showed no evidence for genuine mutant molecules, placing an upper limit of insertion and deletion rates of 2610 77 and 3610 77 , respectively, in the individuals tested. Such re-arrangements at these loci therefore occur at a rate lower than that detectable by the most sensitive methods currently available.

The highly variable human minisatellites MS32 (D1S8), MS31A (D7S21), and CEB1 (D2S90) all show recombination-based repeat instability restricted to the germline. Mutation usually results in polar interallelic conversion or occasionally in crossovers, which, at MS32 at least, extend into DNA flanking the repeat array, defining a localized recombination hotspot and suggesting that cis-acting elements in flanking DNA can influence repeat instability. Therefore, comparative sequence analysis was performed to search for common flanking elements associated with these unstable loci. All three minisatellites are located in GC-rich DNA abundant in dispersed and tandem repetitive elements. There were no significant sequence similarities between different loci upstream of the unstable end of the repeat array. Only one of the three loci showed clear evidence for putative coding sequences near the minisatellite. No consistent patterns of thermal stability or DNA secondary structure were shared by DNA flanking these loci. This work extends previous data on the genomic environment of minisatellites. In addition, this work suggests that recombinational activity is not controlled by primary or secondary characteristics of the DNA sequence flanking the repeat array and is not obviously associated with gene promoters as seen in yeast.[The sequence data described in this paper have been submitted to the GenBank data library under accession nos. AF048727(CEB1), AF048728 (MS31A), and AF048729 (MS32).]