Table 3.

Comparison of two simple tiling metrics that incorporate repetitive nucleotides to improve non-repetitive sequence coverage

			Case 1: Threshold repeat inclusion (50 bp)			Case 2: Percentage repeat inclusion (25%)
Organism	Genome size	Percent repeats	Percent non-repeat bp covered	Percent repeat bp included vs all non-repeat bp	Tile quality	Percent non-repeat bp covered	Percent repeat bp included vs all non-repeat bp	Tile quality
Pan troglodytes	3,083,993,401	57.74	64.85	4.15	62.04	66.85	17.94	52.24
Homo sapiens	3,070,537,687	52.38	65.01	4.09	62.24	67.11	18.22	52.16
Rattus norvegicus	2,795,745,218	48.75	66.66	4.28	63.68	69.42	19.84	52.24
Mus musculus	2,638,213,512	45.62	77.56	4.30	74.07	80.82	20.15	60.43
Caenorhabditis elegans	100,277,879	11.26	89.71	2.18	96.68	99.84	11.12	87.47
Drosophila melanogaster	129,323,838	14.23	97.63	0.03	99.97	100	2.39	97.55
Fugu rubripes	349,519,338	15.06	95.09	1.86	97.74	100	6.33	93.24
Arabidopsis thaliana	119,186,497	0.16	99.51	1.29	98.22	100	13.29	84.68

In Case 1, repeat sequences ≤50 bp were allowed, and in Case 2 up to 25% of a tile may contain repetitive nucleotides. As in Table 1, tile sizes range from 300 bp to 1.5 kb. Case 1 achieves only marginal improvement in non-repetitive sequence coverage when compared with the same level of repeat nucleotide inclusion in the optimal tiling case. Non-repetitive sequence coverage in mammalian genomes falls sharply in Case 2 despite the inclusion of a high percentage of repetitive DNA. In each case, performance on mammalian genomes is significantly lower than that of the optimal tiling algorithm (Table 2).