To date, most published analyses of genomic rearrangements using microarrays have utilized arrays which sample the genome with one clone of 100- to 200-kb in length every 1 Mb (14,22,23). Although these arrays greatly improve on the resolution of analyses possible with metaphase chromosomes as targets for hybridization, higher resolution arrays would allow more detailed analysis of rearrangements. In particular, we are interested in the molecular mechanisms involved in the generation of the rearrangements we identify and for this we require DNA sequence across the breakpoint regions. The sequence from junction regions is most efficiently and cost-effectively generated from PCR or long range PCR products for which primers need to be designed to within a few kilobases of the breakpoints, thus, requiring high resolution analysis of breakpoint position.
One way of improving resolution is to increase the density of clones on the array. Several groups have generated arrays of overlapping clones to targeted regions or whole chromosomes (9,24-28). We have used an array of overlapping clones for the whole of chromosome 1 to analyze patients with constitutional 1p36 deletion syndrome (29). We found two patients where the first had a deletion restricted to the most terminal 2.5 Mb of 1p36.33, whereas the second had a deletion of 6.9 Mb in length, starting 3 Mb from the terminal region. As these two patients showed overlapping phenotypes but nonoverlapping 1p36 deletions, we believe that 1p36 deletion syndrome may be caused by positional effects as well as more conventional contiguous gene deletion. Recently, the first whole genome, bacterial artificial chromosome (BAC) tiling path array, comprising approx 32,000 fingerprint-mapped RPCI-11 BAC clones, has been described (30). Although this array is initially being applied to solid tumors (31), it is clear that such arrays would similarly improve the resolution of the analysis of constitutional rearrangements. In particular, the largely overlapping nature of the clone inserts ensures that breakpoints will be identified to within a clone. Our own genome tiling path array of more than 30,000 clones largely selected from the clones used to generate the reference human sequence (the Golden Path) is in validation and we intend to implement this array as our initial tool for the analysis of constitutional rearrangements.
A second way of improving array resolution is to reduce the length of the genomic sequences spotted onto the arrays. For a first increase in resolution, we routinely utilize fosmid libraries whose clones have been mapped onto the reference sequence. These clones show a high degree of overlap (see Fig. 5) and this redundancy can be used to improve the resolution of the array even beyond the size of the clones as was originally described by Albertson (32). Figure 6 shows a microdeletion on chromosome 9 analyzed using a 1-Mb array and with a custom high density fosmid array targeted to the deletion breakpoints. With such fosmid arrays breakpoints can be positioned to within 20-40 kb.
Clones with inserts of even smaller size are also useful for breakpoint mapping. We have utilized small insert libraries from flow-sorted chromosomes, which have been sequenced in an effort to identify single nucleotide polymorphisms in the human genome. Although these libraries only have a two- to threefold coverage of each chromosome, their small insert size compensates for occasional gaps in tile paths. Figure 7 shows aCGH profiles delineating the proximal breakpoint of the previously shown chromosome 9 microdeletion, using the 1-Mb array, the custom high density fosmid array and finally a high density small insert clones array (1 to 4 kb in length). Custom arrays of this type enable breakpoints to be mapped to well within the distance required for long-range PCR products to be generated from junction regions and sequencing of breakpoints.
Was this article helpful?