Plant and animal genomes are currently sequenced either by a global shotgun sequencing approach  or by sequencing of large-insert genomic clones and assembling the global genome sequence from them (ordered-clone approach) . The former approach is inherently faster and more economical since the entire genome sequence is generated in a single operation. To assemble a genome sequence, it is necessary to identify overlaps of individual reads among vast numbers of other reads. The presence of repeated sequences among the reads makes this task challenging in some genomes. This aspect of genome architecture is greatly exacerbated in plants with large genomes by the precipitous turnover of repeated sequences in the intergenic spaces. For instance, in the tribe Triticeae of the grass family, in which the sizes of genomes in diploid species range from 3.3 to 8.1 Gbp (reviewed in ), sequences filling the intergenic space are almost entirely replaced in about 3 million years, which is a turnover rate orders of magnitude faster than in primate genomes . Because of large genome size and fast turnover rate of repeated sequences, the Triticeae genomes contain large numbers of very similar nucleotide sequences, which has precluded the use of the shotgun genome sequencing approach for diploid Triticeae species.
A special challenge presented to genome sequencing in plants is polyploidy. A large percentage of seed plants are polyploid . Probably all plants are ancient polyploids (paleopolyploids) but since paleopolyploidy does not usually complicate genome sequencing, paleopolyploidy is not considered in this study. Plant polyploids are categorized as either autopolyploids with identical genomes or allopolyploids with related genomes that were contributed by different diploid species. A vast majority of plant polyploids are allopolyploids. The need to allocate sequence reads to respective genomes makes it exceedingly difficult to assemble global genome sequences of polyploid species from whole-genome shotgun sequence reads. For that reason, no polyploid plant genome has yet been sequenced by this approach.
The alternative approach, based on sequencing large-insert clones, potentially avoids the factors limiting the shotgun sequencing approach. The advent of the high-information-content-fingerprinting (HICF) of bacterial artificial chromosome (BAC) clones greatly increased fingerprinting throughput and fidelity [6–8]. With the five-color SNaPshot HICF technology , computer-driven fingerprint editing , contig assembly with the FPC program [10, 11], and contig anchoring on high-resolution genetic maps with the highly multiplexed Illumina GoldenGate™ assays , it is now theoretically possible to construct physical maps for most diploid plants and animals, including ancient polyploids, such as maize and soybean [13, 14].
The SNaPshot HICF fingerprinting technology is based on restriction digestion of the DNA of each BAC clone by multiple restriction endonucleases and sizing a portion of the fragments with capillary electrophoresis. Contigs are then assembled on the basis of shared portions of the restriction profiles of the BAC clones. It has been tacitly assumed that BAC clones from homoeologous chromosome regions in an allopolyploid will have too many restriction fragments in common and will be included into single contigs during contig assembly. Consequently, physical mapping based on the SNaPshot HICF technology has not been pursued to any significant extent in recently evolved allopolyploids, with the sole exception of hexaploid wheat, Triticum aestivum.
Polyploid wheat species of economical importance are either allotetraploid (T. turgidum, genome formula AABB) or allohexaploid (T. aestivum, genome formula AABBDD). The A, B, and D genomes were contributed by three different diploid species which radiated from a common ancestor between 2.5 and 4.5 million years ago, depending on which of several estimates is used, and are approximately equally diverged from each other at the molecular level [15, 16]. Because of the recent divergence of the three ancestors, it was assumed that the assembly of contigs from a global T. aestivum BAC library would not produce physical maps of wheat chromosomes that would be of adequate quality for genome sequencing. Instead, technological advances in flow-sorting of chromosomes and the unique availability of individual chromosome and chromosome arm genetic stocks for wheat suggested an alternative procedure for generating hexaploid wheat physical maps. Chromosome-specific or chromosome-arm-specific BAC libraries are being constructed from DNA produced by flow-sorting of complete and telocentric chromosomes of T. aestivum  and used for the construction of the physical maps of the 21 T. aestivum chromosomes http://www.wheatgenome.org/. Such BAC libraries are easier to handle and simplify contig assembly compared to a global T. aestivum BAC library comprising over one million clones. Their availability also facilitates division of labor and international collaboration on the development of wheat sequence-ready physical maps. The successful construction of the physical map of T. aestivum chromosome 3B  from a chromosome 3B BAC library demonstrated the feasibility of this approach.
For chromosome flow-sorting to be a general approach to produce physical maps of allopolyploid species, each chromosome in the karyotype of a targeted allopolyploid would have to be a unique size - a condition, which is seldom met. Alternatively, special cytogenetic stocks, such as telosomic lines and chromosome addition lines must be available for all chromosome arms present in the genome or developed de novo . For these reasons, the physical mapping strategy adopted for sequencing of T. aestivum is not generally applicable, and genomes of most polyploid plants, including tetraploid wheat, cannot be physically mapped by this approach.
An assessment of the utility of the current HICF technology for the construction of physical maps of polyploid plants from global BAC libraries is therefore of central importance for advancing genome research on polyploid organisms. To date, because of the large costs involved, no relevant data on this subject exists. However, the ability to produce chromosome-specific physical maps in hexaploid wheat provides the opportunity to undertake the assessment using a single set of homoeologous chromosome arms. Such an assessment is reported here.
BAC libraries constructed from flow-sorted telocentric chromosomes 3AS and 3DS and complete chromosome 3B were employed. Telosomes 3AS and 3DS are homoeologous to each other and both are homoeologous to the short arm of chromosome 3B (arm 3BS). The three libraries were fingerprinted. Contigs were either assembled from the clones of a single library or fingerprints were merged, and contigs were assembled from the clones of the merged library. Since the origin of each clone in the merged library was known, the frequency of inclusion of clones from more than a single chromosome arm into the contigs could be quantified for the entire population of contigs.