RNA-protein binding motifs mining with a new hybrid deep learning based cross-domain knowledge integration approach

Pan, Xiaoyong; Shen, Hong-Bin

doi:10.1186/s12859-017-1561-8

Table 1 The AUC performance comparison between iDeep and other methods on 31 experiments

From: RNA-protein binding motifs mining with a new hybrid deep learning based cross-domain knowledge integration approach

Protein	iDeep	iONMF	NMF	SNMF	QNO	Oli	iDeep-kmer	DeepBind
1 Ago/EIF	0.90	0.89	0.89	0.85	0.87	0.61	0.87	0.69
2 Ago2-MNase	0.73	0.71	0.69	0.66	0.69	0.51	0.67	0.53
3 Ago2-1	0.91	0.81	0.81	0.76	0.83	0.80	0.82	0.81
4 Ago2-2	0.91	0.84	0.82	0.79	0.82	0.80	0.83	0.81
5 Ago2	0.74	0.73	0.71	0.65	0.66	0.53	0.65	0.58
6 eIF4AIII-1	0.94	0.92	0.91	0.78	0.95	0.92	0.95	0.93
7 eIF4AIII-2	0.97	0.93	0.93	0.67	0.64	0.93	0.94	0.93
8 ELAVL1-1	0.96	0.91	0.89	0.71	0.80	0.89	0.95	0.90
9 ELAVL1-MNase	0.68	0.71	0.70	0.68	0.70	0.49	0.66	0.54
10 ELAVL1A	0.94	0.94	0.93	0.91	0.92	0.84	0.95	0.87
11 ELAVL1-2	0.97	0.95	0.94	0.90	0.95	0.88	0.97	0.91
12 ESWR1	0.95	0.87	0.85	0.80	0.85	0.81	0.92	0.88
13 FUS	0.92	0.81	0.73	0.55	0.65	0. 85	0.87	0.92
14 Mut FUS	0.97	0.96	0.95	0.91	0.94	0.82	0.97	0.91
15 IGFBP1-3	0.95	0.93	0.92	0.89	0.91	0.57	0.93	0.68
16 hnRNPC-1	0.93	0.95	0.93	0.45	0.63	0.88	0.92	0.95
17 hnRNPC-2	0.97	0.97	0.96	0.48	0.70	0.94	0.95	0.97
18 hnRNPL-1	0.82	0.74	0.73	0.70	0.77	0.39	0.79	0.76
19 hnRNPL-2	0.82	0.66	0.62	0.56	0.61	0.47	0.72	0.74
20 hnRNPL-like	0.79	0.69	0.67	0.63	0.68	0.56	0.70	0.70
21 MOV10	0.97	0.96	0.96	0.89	0.92	0.78	0.97	0.80
22 Nsun2	0.87	0.81	0.80	0.69	0.82	0.75	0.81	0.84
23 PUM2	0.98	0.93	0.92	0.86	0.89	0.94	0.98	0.93
24 QKI	0.95	0.84	0.77	0.52	0.62	0.92	0.92	0.95
25 SRSF1	0.92	0.85	0.85	0.73	0.86	0.84	0.85	0.85
26 TAF15	0.97	0.91	0.89	0.82	0.91	0.80	0.95	0.95
27 TDP-43	0.89	0.84	0.78	0.45	0.57	0.88	0.85	0.89
28 TIA1	0.94	0.93	0.92	0.86	0.90	0.84	0.96	0.90
29 TIAL1	0.92	0.87	0.86	0.73	0.85	0.83	0.90	0.87
30 U2AF2	0.95	0.82	0.74	0.61	0.70	0.86	0.91	0.95
31 U2AF2(KD)	0.92	0.80	0.74	0.60	0.74	0.84	0.88	0.91
Mean	0.90 ±0.08	0.85 ±0.08	0.83 ±0.10	0.71 ±0.14	0.79 ±0.12	0.77 ±0.16	0.87 ±0.09	0.83±0.12

The performance of iONMF, NMF, SNMF and QNO are taken from [5]. DeepBind, Oli and iDeep-kmer perform on the same data with iDeep, and iDeep-kmer used kmer to replace CNN sequence and motif modalities in iDeep
The boldface indicates this performance is the best among the compared methods

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com

BMC Bioinformatics

Contact us