Assembling ordered, overlapping sets (contigs) of high-quality, - sequence-ready clones has long been considered an essential step toward human genome sequencing. Not only do the clones provide uniform materials for sequencing, but, because they have been mapped to precise genomic locations, the DNA sequence obtained from them can be located on the chromosomes with minimal uncertainty. Very useful low-resolution maps have been produced by several methods. [See, for example, summary maps at the MIT Whitehead Institute (http://www.broad.mit.edu/) and T.J. Hudson et al., Science 270, 1945-54.] However, only 3 of 24 chromosomes (16, 19, and 22) are substantially covered by contigs of sequence-ready clones.

The availability of newer, more stable clone resources containing large human DNA inserts (up to 150,000 bases) has stimulated an alternative strategy to contig building for complete genome - sequencing. In this new approach, described in Nature (381, 364-66), about 500 bases of sequence are obtained from both ends of BAC or PAC clone inserts. BAC-PAC end sequencing is performed on clones from a deep (about 20-fold), arrayed library, and all clone names, end sequences, and other useful data are entered into a public database. When a given clone is sequenced, any researcher can search for and identify additional clones with overlapping - sequences. These "hits" can then guide the choice of the next overlapping BAC or PAC clone to be sequenced. In this way, contig building (determining overlapping pieces and ordering all the clones) is bypassed; it happens as a consequence of sequencing, not as a prerequisite.

Proponents of this strategy assert that it will be a simpler, faster, and cheaper way to obtain the total genome sequence the ultimate goal of the Human Genome Project. Also, because clone resources and their end-sequence data can be made available worldwide, this approach will support geographically dispersed participation in genome sequencing as well as easier access to clones for other genome-related purposes.

The general BAC-PAC end-sequencing strategy was discussed at several recent meetings, including a December 1995 DOE-sponsored BAC workshop and the February 1996 Bermuda conference on high-throughput sequencing sponsored largely by the Wellcome Trust [see HGN 7(6), 19]. This past summer, DOE reviewed end-sequencing applications and made 2-year awards to the following teams to undertake pilot projects for testing the feasibility of this strategy, its technologies, and its economics:

* Mark Adams [The Institute for Genomic Research (TIGR)], Leroy Hood (University of Washington, Seattle), and Melvin Simon (California Institute of Technology); and
* Glen Evans with Harold Garner [University of Texas Southwest (UTSW) Medical Center], Pieter de Jong (Roswell Park Cancer Center), and Julie Korenberg (Cedars Sinai Medical Center).