Table 1.

AF4 for simulated datasets (Liu et al., 2016)

Samples		AF4								ChimeRScope	STAR-Fusion
		Targeted				De novo
Length (bp)	Coverage	TP	FP	FN	F-score	TP	FP	FN	F-score	F-score	F-score
50	5×	146	0	4	0.986	143	1	7	0.973	0.948	0.416
50	20×	147	0	3	0.990	144	2	6	0.973	0.954	0.855
50	50×	147	0	3	0.990	144	2	6	0.973	0.947	0.867
50	100×	147	0	3	0.990	144	3	6	0.970	0.908	0.868
50	200×	147	0	3	0.990	144	3	6	0.970	0.905	0.875
75	5×	144	0	6	0.980	141	1	9	0.966	0.948	0.700
75	20×	147	0	3	0.990	143	1	7	0.973	0.949	0.865
75	50×	147	0	3	0.990	143	1	7	0.973	0.957	0.872
75	100×	147	0	3	0.990	144	2	6	0.973	0.954	0.875
75	200×	147	0	3	0.990	144	3	6	0.970	0.947	0.875
100	5×	144	0	6	0.980	140	0	10	0.966	0.940	0.692
100	20×	147	0	3	0.990	143	1	7	0.973	0.957	0.878
100	50×	147	0	3	0.990	143	1	7	0.973	0.957	0.875
100	100×	147	0	3	0.990	143	1	7	0.973	0.957	0.872
100	200×	147	0	3	0.990	143	1	7	0.973	0.957	0.877

Note: Each dataset contains fused reads of 150 gene pairs. In the targeted mode, the names of the 655 target gene pairs (including all 150 target ones) are given in the command line, along with the transcriptome. In the de novo mode, only the transcriptome is given.