Table 2.

Phase determinations as inferred from the simulated assemblies for six alternative sequencing strategies. The greater N₅₀ and % sequenced fractions identify those strategies with longer contigs and thereby better phase determinations (Figure 4). All simulations were done with two copies of chromosome 2. Heterozygous sites were modeled as single base differences with a rate of 0.1% per site. The unpaired reads were of length 400 or 2000 bases. The reads of the mate pairs were 100 bases long with intervening distances from a normal distribution between 1500 and 2500 bp. All shotgun simulations were performed with 50-times total depth (i.e., 25-times coverage for each homolog of this diploid). Each BAC pooling consisted of 243 pools with 100 clones apiece. This pooling resulted in 10-times BAC coverage of the two reference homologs and five-times read coverage of each clone for the same overall depth of 50 times.

Strategy	Data		N50 / bp	Sequenced fraction
Shotgun sequencing	400-base unpaired reads	734		52.47%
	2000-base unpaired reads	9,311		95.65%
	2 x 100-base paired reads	2,274		57.66%
BAC pooling	400-base unpaired reads	384,621		97.03%
	2000-base unpaired reads	1,594,313		97.17%
	2 x 100-base paired reads	2,932,847		97.17%