C O M P U T a T I O N a L B I O L O G Y

33 T he prevailing method of determining the sequence of a long DNA segment is the shotgun sequencing approach, in which a random sampling of short fragment sequences is acquired and then assembled by a computer program to infer the sampled segment's sequence. In the early 1980s, such segments were typically on the order of 5,000 to 10,000 base pairs (5 to 10 kbp). By 1990, this method was sequencing segments on the order of 40 kbp, and by 1995, the entire 1,800-kbp H. Influenzae bacterium had been sequenced. 1 The source segment is clearly becoming extremely large without any concomitant increase in the length of the sampled fragment sequences. Soon, shotgun data sets could well consist of millions of sampled fragment sequences and will require significant computational resources to assemble. The whole-genome shotgun sequencing of H. Influenzae in 1995 showed that direct shotgun se-quencing could handle a much larger source sequence segment than biologists had commonly thought. Before that, cosmid-sized clones of 30 to 50 kbps were considered this approach's upper limit. (A cosmid is a type of vector for manipulating and replicating inserted pieces of DNA.) Now, the shotgun sequencing of 200 kbp bacterial artificial chromosomes (BACs) is a given. This achievement inspired Jim Weber and me to propose the use of a shotgun approach to sequence the human genome, 2 after which we requested funding for a pilot project from the US National Institutes of Health. The established community rejected our controversial proposal, 3 but in May of 1998, Craig Venter and the Perkin-Elmer Corporation announced a new private venture, Celera Genomics, aimed at using a whole-genome shotgun approach to sequence the fruit fly Drosophila (≈120 Mbp) in 1999 and the human genome (≈3.5 Gbp) by 2001. 4 After introducing the basic technology of shotgun DNA sequencing and briefly summarizing the computational and algorithmic results to date on the problem of assembling shotgun data sets, this article characterizes the nature of DNA sequences and improvements in sequenc-ing technology that affect the computational problem. It then analyzes three current proposals for sequencing the human genome, including the one we're pursuing at Celera Genomics. Shotgun DNA sequencing basics In the late 1970s, the first procedures emerged for determining the sequence of nucleotides WHOLE-GENOME DNA SEQUENCING Computation is integrally and powerfully involved with the DNA sequencing technology that promises to reveal the complete human DNA sequence in the next several …