INSnet: a method for detecting insertions based on deep learning network

BMC Bioinformatics

Table 4 The performance of insertions in different sizes on 69× data about HG002

Phase	INSnet	cuteSV	SVIM	Sniffles
CLR69X
50–200
Precision	0.8961	0.8639	0.8958	0.3106
Recall	0.8817	0.9458	0.9421	0.9126
F1	0.8888	0.903	0.9184	0.463
200–500
Precision	0.9337	0.9506	0.6964	0.9108
Recall	0.9173	0.9155	0.8803	0.9525
F1	0.9254	0.9327	0.7776	0.9312
500–1000
Precision	0.8276	0.9207	0.9226	0.8617
Recall	0.8484	0.7626	0.7222	0.8182
F1	0.8379	0.8343	0.8102	0.8394
1000–5000
Precision	0.9142	0.95	0.9592	0.2327
Recall	0.9142	0.6524	0.2017	0.4893
F1	0.9142	0.7735	0.3333	0.3154
5000-
Precision	0.8888	1	0	0.0399
Recall	0.5	0.1458	0	0.2292
F1	0.64	0.2545	0	0.0679

ISSN: 1471-2105