The proportion of the 19,667 RefSeq mRNA sequences that can be aligned to each of the genomes at various coverage thresholds and more than 95% sequence identity.

Dot-plot representation of sample assembly comparison results. Horizontal axes correspond to intervals along NCBI-34, and vertical axes correspond to intervals along various assemblies, with the sequences starting from the bottom left corner. Diagonal lines show the relative positions and orientations of matches. Identical sequences would yield one diagonal line. Vertical bars represent gaps between NCBI-34 contigs. Selected regions were chosen to represent general observations regarding the assemblies; related figures of entire chromosomes are provided for all chromosomes in Data Set 7. (a) Illustration of a region in which WGSA can augment NCBI-34. Shown are the first 6 Mbp of NCBI-34 human chromosome 1 versus part of a single scaffold of WGSA. The second NCBI-34 contig is inverted, and the third and fourth contigs are interchanged, compared with WGSA. We postulate that this is an NCBI-34 contig mapping problem. Alternative explanations, such as misassembly or polymorphisms within the WGSA scaffold that coincidentally occur at the boundaries of NCBI-34 contigs, are improbable. (b–f) Comparison of the NCBI-34 human chromosome 1 region from 34–40 Mbp against the primary matching regions of WGSA (b), WGA (c), CSA (d), HG06 (e), and NCBI-28 (f). (See main text for description of assemblies.) WGSA agrees closely with NCBI-34 and spans and largely fills two gaps between NCBI-34 contigs. All other assemblies have multiple order and orientation errors. For all but HG06, the misplaced segments correspond to entire scaffolds (data not shown). For HG06, errors are a mix of within-scaffold rearrangements and scaffold order and orientation. WGA and HG06 both have a relatively large number of small, misplaced scaffolds, whereas CSA and NCBI-28 have a few, larger scaffolds that are misplaced.