Unsupervised Two-Way Clustering of Metagenomic Sequences

Table 1

Performance of Gaussian mixture model (without word grouping) on datasets containing more than 2 species, at various abundances on reads of length 500 bp. AR stands for abundance ratio.

Species	AR	Number reads	Accuracy (%)
T. thermophilis	1	50000
A. vinelandii	3		87.51
N. meningitidis	2
E. coli 536	1	50000
S. acidocaldarius	2		97.01
H. salinarum R1	2
C. jejuni RM1221	3	60000
H. salinarium R1	2		96.61
E. coli	1
P. horikoshii OT3	3
S. erythraea	1	60000
M. thermoautotrophicum	1		90.28
B. burgdorferi ZS7	1
E. coli 536	1
B. burgdorferi ZS7	1	75000
C. jejuni RM1221	1		85.04
E. coli 536	1
H. salinarum R1	1
P. horikoshii OT3	1