Metagenome Fragment Classification Using 𝑁-Mer Frequency Profiles

Table 1

Comparison of run times for 𝑁=10 (general and optimized methods) and 𝑁=100 (general method only). Simulations are run on one 2 × 2.0 GHz Intel dual-core Xeon, 2 GB RAM, and 2 × 80 GB HD. Many temporary files are needed for the sort process and are saved to a 3 TB RAID drive which is connected through a cluster-head terminal machine.

Name	Genome size (Mbp)	Optimized alg. 𝑁=10 (min)	General alg. 𝑁=10 (min)	General alg. 𝑁=100 (min)
Human	3142.05	2.4	212.7	1877.2
Zebrafish	1578.26	1.1	106.25	506.5
Fruitfly	135.25	0.1	6.5	26.4
E. coli K12	4.7	0.01	0.18	0.63