Randomness Versus Order

M i l e s to n e s

DOI: 10.1038/nrg2245 M iles tone 1 0 Randomness versus order

Whereas randomness is avoided used — in their opinion mistakenly The H. influenzae genome, in most experimental techniques, — direct sequencing strategies to however, was a mere DNA fragment it is fundamental to sequencing finish compared with the 1,500-fold longer approaches. In the race to sequence the last 10% of the bacteriophage λ ~3 billion base-pair human genome. the human genome, research groups sequence. In 1991, Al Edwards and In 1996, Craig Venter and colleagues had to choose between the random Thomas Caskey proposed a method proposed that the whole-genome whole-genome shotgun sequencing to maximize efficiency by minimiz- shotgun approach could be used to approach or the more ordered map- ing gap formation and redundancy: sequence the human genome owing based sequencing approach. sequence both ends (but not the to two factors: its past successes When Frederick Sanger and middle) of a long clone, rather than in assembling genomes and the colleagues sequenced the 48-kb the entirety of a short clone. development of bacterial artificial bacteriophage λ genome in 1982, Although the shotgun approach chromosomes (BAC) libraries, which the community was still undecided was now accepted for sequencing allowed large fragments of DNA to as to whether directed or random short stretches of DNA, map-based be cloned. sequencing strategies were better. techniques were still considered A showdown ensued, with the With directed strategies, DNA necessary for large genomes. Like biotechnology firm Celera Genomics sequences were broken down into the directed strategies, map-based wielding whole-genome shotgun ordered and overlapping fragments sequencing subdivided the genome sequencing and the International to build a map of the genome, and into ordered 40-kb fragments, which Human Genome Sequencing these fragments were then cloned were then sequenced using the Consortium wielding map-based and sequenced. With the shotgun shotgun approach. In 1995, however, sequencing. Yet, when the dust set- approach, DNA sequences were bro- Robert Fleischmann and colleagues tled, it was a draw — both groups ken randomly, cloned, sequenced and used a whole-genome shotgun published their initial drafts of the then pieced together by analysing approach to sequence the 1,800-kb human genome concurrently in 2001. the overlap. Sanger et al. compared genome of Haemophilus influenzae Asher Mullard, Assistant Editor, these strategies while sequencing — the first complete genome of a Nature Reviews Microbiology and bacteriophage λ and reported that free-living organism. The authors Nature Reviews Molecular Cell Biology the random approach was faster than had randomly generated large 40-kb ORIGINAL RESEARCH PAPERS Sanger, F. et al. any directed method. fragments and had thereby bypassed Nucleotide sequence of bacteriophage λ DNA. One problem with the random the mapping stage. In doing so, they J. Mol. Biol. 162, 729–773 (1982) | Edwards, A. & approach, however, was that of filling had proved that genome-assembly Caskey, C. T. Closure strategies for random DNA sequencing. Methods 3, 41–47 (1991) | gaps when the sequence was nearly programmes that matched overlap Fleischmann, R. D. et al. Whole-genome random complete (or closure), as randomly were reliable and that whole-genome sequencing and assembly of Haemophilus influenzae Rd. Science 269, 496–512 (1995) | selected clones were often shotgun sequencing worked, in Venter, J. C. et al. A new strategy for genome redundant. For instance, Sanger et al. principle. sequencing. Nature 381, 364–366 (1996)

In the race to sequence the human genome, research groups had to choose between the random whole-genome shotgun sequencing approach or the more ordered map-based sequencing approach.