Table 4.

Performance analysis of six k-spectrum-based error correctors as evaluated using six synthetic Illumina datasets

Dataset	Method	TP	FP	FN	Recall	Gain	Precision	F-score
EC-1	Reptile	2335361	144751	451889	0.8378	0.7859	0.9416	0.8867
36 bp	Lighter	2695425	72843	91825	0.9671	0.9409	0.9737	0.9704
70×	Bless	2624659	48342	56279	0.9790	0.9610	0.9819	0.9805
k = 19	Bloocoo	2411701	22259	375549	0.8653	0.8573	0.9908	0.9238
	Musket	2701885	61096	85365	0.9694	0.9474	0.9779	0.9736
	Trowel	1246340	705438	1539825	0.4473	0.1941	0.6386	0.5261
EC-2	Reptile	681551	140039	114910	0.8557	0.6799	0.8296	0.8424
36 bp	Lighter	108241	58579	688220	0.1359	0.0624	0.6488	0.2247
20×	Bless	779824	18095	16637	0.9791	0.9564	0.9773	0.9782
k = 17	Bloocoo	689322	6454	107139	0.8655	0.8574	0.9907	0.9239
	Musket	767087	18182	29374	0.9631	0.9403	0.9768	0.9699
	Trowel	434885	19167	361576	0.5460	0.5220	0.9578	0.6955
EC-3	Reptile	105	461	876053	0.0001	-0.0004	0.1855	0.0002
100 bp	Lighter	858125	2446	18033	0.9794	0.9766	0.9972	0.9882
20×	Bless	746	872860	875412	0.0008	-0.9954	0.0009	0.0009
k = 24	Bloocoo	79790	3644539	796368	0.0911	-4.0686	0.0214	0.0347
	Musket	873592	1645	2566	0.9971	0.9952	0.9981	0.9976
	Trowel	155	178354	876003	0.0002	-0.2034	0.0009	0.0003
BC-1	Reptile	382043	22303	16602	0.9584	0.9024	0.9448	0.9515
56 bp	Lighter	331759	15470	141618	0.7008	0.6682	0.9554	0.8086
50×	Bless	429017	34018	11943	0.9729	0.8958	0.9265	0.9492
k = 27	Bloocoo	410156	24127	63221	0.8664	0.8155	0.9444	0.9038
	Musket	355015	47460	118362	0.7500	0.6497	0.8821	0.8107
	Trowel	55277	4976	26744	0.6739	0.6133	0.9174	0.7770
BC-2	Reptile	497425	116	208081	0.7051	0.7049	0.9998	0.8269
100 bp	Lighter	698089	159	7417	0.9895	0.9893	0.9998	0.9946
120×	Bless	–	–	–	–	–	–	–
k = 31	Bloocoo	27409	1278837	678097	0.0389	-1.7738	0.0210	0.0272
	Musket	703882	68	1624	0.9977	0.9976	0.9999	0.9988
	Trowel	652845	108	52661	0.9254	0.9252	0.9998	0.9612
DM	Reptile	11702183	187733	517322	0.9577	0.9423	0.9842	0.9708
100 bp	Lighter	42	23055867	12224293	0.0000	-1.8861	0.0000	0.0000
10×	Bless	11122683	126388	1101652	0.9099	0.8995	0.9888	0.9477
k = 21	Bloocoo	–	–	–	–	–	–	–
	Musket	11550483	163838	673852	0.9449	0.9315	0.9860	0.9650
	Trowel	1197127	384403	11027208	0.0979	0.0665	0.7569	0.1734

In the first column, dataset ID, read length, genome coverage, and the optimal k estimated using KmerGenie are shown. The values in TP, FP, and FN columns are numbers of bases. Italicized values denote the best performer with regard to a specific evaluation measure for a dataset. The symbol “–” indicates that a method failed to process a specific dataset