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

Table 5 Comparison of the alignment-free distances and the benchmark MSA distance for 70 Gammaproteobacteria genomes

		d ^NCD	d	d ^CDM	CVTree	$d_{2}^{S}$	co-phylog
	parsimony score	17	18	18	17	18	25
16s rRNA sequences	tree symmetric difference	50	52	52	50	62	108
	distance correlation	0.93	0.90	0.93	0.92	0.92	0.65
	parsimony score	22	22	21	21	31	26
Genome sequences	tree symmetric difference	80	78	76	84	110	110
	distance correlation	0.47	0.46	0.47	0.67	0.50	0.45
	parsimony score	21	19	23	24	32	28
NGS short reads	tree symmetric difference	90	70	84	88	114	116
	distance correlation	0.60	0.58	0.53	0.63	0.48	0.42

The NGS short reads were simulated from the whole genome sequences using the Exact model of MetaSim at 1 × sampling depth. The two smallest parsimony scores, the two smallest tree symmetric differences and the two highest correlation coefficients are highlighted in boldface. For CVTree, we used k = 7 for the 16S rRNA data set and k = 12 for the whole genome and NGS data sets. For d 2S, we used k = 6 for the 16S rRNA data set and k = 8 for the whole genome and NGS data sets.

ISSN: 1756-0500