DeepCRBP: improved predicting function of circRNA-RBP binding sites with deep feature learning

doi:10.1007/s11704-023-2798-1

Front. Comput. Sci.

2024, Vol. 18

Issue (2) : 182907 https://doi.org/10.1007/s11704-023-2798-1

Interdisciplinary

DeepCRBP: improved predicting function of circRNA-RBP binding sites with deep feature learning

Zishan XU, Linlin SONG, Shichao LIU(

), Wen ZHANG(

)

College of Informatics, Huazhong Agricultural University, Wuhan 430070, China

Download: PDF(1329 KB) HTML
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks

Corresponding Author(s): Shichao LIU,Wen ZHANG

Just Accepted Date: 05 June 2023 Issue Date: 19 July 2023

Cite this article:

Zishan XU,Linlin SONG,Shichao LIU, et al. DeepCRBP: improved predicting function of circRNA-RBP binding sites with deep feature learning[J]. Front. Comput. Sci., 2024, 18(2): 182907.

URL:

https://academic.hep.com.cn/fcs/EN/10.1007/s11704-023-2798-1
https://academic.hep.com.cn/fcs/EN/Y2024/V18/I2/182907

Fig.1 Overview of DeepCRBP. DeepCRBP is composed of spatial encoding module, sequence encoding module and prediction module

	DeepCRBP		DMSK		DHN		Debcan		CRIP
	AUC	ACC	AUC	ACC	AUC	ACC	AUC	ACC	AUC	ACC
AGO1	0.966	0.910	0.906	0.829	0.899	0.812	0.866	0.786	0.903	0.824
AGO2	0.928	0.858	0.802	0.723	0.837	0.738	0.836	0.756	0.804	0.728
AGO3	0.965	0.904	0.930	0.834	0.941	0.857	0.872	0.803	0.906	0.815
ALKBH5	0.981	0.928	0.924	0.841	0.992	0.943	0.923	0.831	0.760	0.690
AUF1	0.993	0.964	0.987	0.947	0.990	0.953	0.963	0.910	0.978	0.945
C17ORF85	0.980	0.935	0.946	0.877	0.990	0.936	0.909	0.838	0.809	0.735
C22ORF28	0.954	0.886	0.897	0.815	0.927	0.837	0.887	0.803	0.867	0.777
CAPRIN1	0.936	0.861	0.857	0.777	0.893	0.801	0.884	0.796	0.839	0.751
DGCR8	0.974	0.926	0.906	0.819	0.930	0.853	0.854	0.773	0.914	0.833
EIF4A3	0.908	0.832	0.803	0.726	0.819	0.732	0.820	0.734	0.816	0.741
EWSR1	0.974	0.913	0.938	0.870	0.960	0.896	0.934	0.861	0.933	0.866
FMRP	0.949	0.880	0.895	0.816	0.911	0.819	0.841	0.761	0.900	0.816
FOX2	0.962	0.907	0.925	0.806	0.978	0.926	0.936	0.873	0.755	0.693
FUS	0.954	0.899	0.853	0.773	0.892	0.807	0.852	0.770	0.854	0.771
FXR1	0.993	0.957	0.983	0.919	0.991	0.958	0.932	0.851	0.964	0.906
FXR2	0.976	0.923	0.946	0.862	0.962	0.893	0.910	0.821	0.934	0.853
HNRNPC	0.989	0.950	0.970	0.920	0.986	0.947	0.969	0.933	0.974	0.920
HUR	0.945	0.877	0.876	0.797	0.893	0.804	0.840	0.759	0.877	0.795
IGF2BP1	0.938	0.866	0.846	0.764	0.875	0.785	0.857	0.777	0.841	0.758
IGF2BP2	0.916	0.838	0.833	0.745	0.852	0.763	0.888	0.809	0.833	0.747
IGF2BP3	0.937	0.869	0.814	0.731	0.840	0.747	0.864	0.784	0.808	0.727
LIN28A	0.963	0.909	0.860	0.772	0.887	0.795	0.866	0.793	0.865	0.778
LIN28B	0.962	0.900	0.891	0.786	0.920	0.838	0.882	0.808	0.886	0.807
METTL3	0.862	0.810	0.869	0.793	0.893	0.786	0.916	0.833	0.819	0.737
MOV10	0.944	0.879	0.848	0.762	0.852	0.771	0.951	0.904	0.833	0.754
PTB	0.955	0.892	0.824	0.741	0.850	0.761	0.869	0.797	0.832	0.747
PUM2	0.985	0.941	0.952	0.888	0.984	0.930	0.963	0.915	0.946	0.881
QKI	0.986	0.949	0.976	0.906	0.997	0.976	0.925	0.852	0.895	0.845
SFRS1	0.988	0.947	0.963	0.889	0.974	0.908	0.887	0.805	0.957	0.887
TAF15	0.993	0.962	0.986	0.934	0.984	0.792	0.923	0.850	0.980	0.943
TDP43	0.977	0.920	0.931	0.855	0.944	0.867	0.964	0.911	0.933	0.855
TIA1	0.986	0.941	0.957	0.891	0.971	0.922	0.964	0.924	0.952	0.868
TIAL1	0.963	0.897	0.908	0.828	0.929	0.843	0.961	0.907	0.909	0.828
TNRC6	0.978	0.931	0.924	0.864	0.977	0.829	0.920	0.840	0.713	0.644
U2AF65	0.967	0.909	0.921	0.842	0.942	0.867	0.934	0.860	0.924	0.848
WTAP	0.942	0.875	0.958	0.894	0.989	0.938	0.916	0.824	0.759	0.649
ZC3H7B	0.943	0.877	0.800	0.720	0.828	0.737	0.896	0.812	0.785	0.709
Avg	0.960	0.903	0.903	0.826	0.926	0.848	0.902	0.829	0.872	0.796

Tab.1 Comparison of DeepCRBP and four compared methods

Fig.2 The average results of ACC and AUC between DeepCRBP and compared methods on 37 circRNA datasets

Fig.3 Results of ablation experiments. Among them, STD and LTD represent short term dependence and long term dependence, respectively

1	J M, Nigro K R, Cho E R, Fearon S E, Kern J M, Ruppert J D, Oliner K W, Kinzler B Vogelstein . Scrambled exons. Cell, 1991, 64( 3): 607–613
2	H, Li Z, Deng H, Yang X, Pan Z, Wei H B, Shen K S, Choi L, Wang S, Wang J Wu . circRNA-binding protein site prediction based on multi-view deep learning, subspace learning and multi-view classifier. Briefings in Bioinformatics, 2022, 23( 1): bbab394
3	K, Zhang X, Pan Y, Yang H B Shen . CRIP: predicting circRNA–RBP-binding sites using a codon-based encoding and hybrid deep neural networks. RNA, 2019, 25( 12): 1604–1615
4	R, Lorenz S H, Bernhart Zu Siederdissen C, Höner H, Tafer C, Flamm P F, Stadler I L Hofacker . ViennaRNA package 2.0. Algorithms for Molecular Biology, 2011, 6: 26
5	Y, Yang Z, Hou Z, Ma X, Li K C Wong . iCircRBP-DHN: identification of circRNA-RBP interaction sites using deep hierarchical network. Briefings in Bioinformatics, 2021, 22( 4): bbaa274

[1]	FCS-22798-OF-ZX_suppl_1	Download
[2]	FCS-22798-OF-ZX_suppl_2	Download

Viewed

Full text

Abstract

Cited

Shared

Discussed