Comparison of next-generation sequencing samples using compression-based distances and its application to phylogenetic reconstruction

Table 4 Comparison of the alignment-free distances and the benchmark co-phylog distance for NGS short reads of 29 Escherichia/Shigella genomes

	d ^NCD	d	d ^CDM	CVTree(k= 9)	CVTree(k= 15)	CVTree(k= 21)	$d_{2}^{S} (k = 8)$
454	16	14	12	22	22	54	50
Exact	12	10	10	22	24	36	50
Empirical	16	12	14	24	18	42	54
Sanger	12	10	10	28	24	40	50
454	0.96	0.96	0.97	0.74	0.87	0.31	0.04
Exact	0.97	0.97	0.98	0.82	0.82	0.61	0.05
Empirical	0.96	0.96	0.96	0.77	0.85	0.38	0.01
Sanger	0.97	0.97	0.97	0.78	0.86	0.36	0.04

The short reads were simulated from the Escherichia/Shigella genomes using four error models 454, Exact, Empirical, and Sanger of the tool MetaSim at 1 × sampling depth. The two smallest tree symmetric differences and the two highest correlation coefficients for each error model are highlighted in boldface. Similar results for 5 × sampling depth can be found in Additional file 1: Table S3.

ISSN: 1756-0500