EnsembleSplice: ensemble deep learning model for splice site prediction

Akpokiro, Victor; Martin, Trevor; Oluwadare, Oluwatosin

doi:10.1186/s12859-022-04971-w

BMC Bioinformatics

Table 3 The cross-validation results for the dataset for the genomic organisms

From: EnsembleSplice: ensemble deep learning model for splice site prediction

Datasets	SpliceSites	Metrics	ENS1	ENS2	ENS3	ENS4	ENS5	ENS6
HS3D	Acceptor	Double fault	0.033	0.00	0.01	0.01	0.007	0.011
		Correlation	0.612	0.06	0.22	0.20	0.21	0.33
		Q-statistics	0.89	0.131	0.50	0.65	0.553	0.83
		Disagreement	0.03	0.00	0.03	0.03	0.02	0.03
		Accuracy	0.89	0.936	0.94	0.93	0.94	0.93
	Donor	Double fault	0.013	0.00	0.00	0.00	0.003	0.003
		Correlation	0.496	0.02	0.18	0.11	0.19	0.20
		Q-Statistics	0.796	− 0.001	0.44	0.37	0.451	0.478
		Disagreement	0.015	0.00	0.01	0.01	0.01	0.01
		Accuracy	0.93	0.958	0.95	0.95	0.94	0.94
A. thaliana	Acceptor	Double fault	0.023	0.003	0.012	0.01	0.011	0.01
		Correlation	0.667	0.215	0.358	0.401	0.413	0.415
		Q-Statistics	0.988	0.713	0.843	0.98	0.982	0.985
		Disagreement	0.023	0.016	0.097	0.027	0.03	0.025
		Accuracy	0.913	0.947	0.946	0.945	0.948	0.942
	Donor	Double fault	0.013	0.019	0.008	0.006	0.007	0.007
		Correlation	0.638	0.132	0.317	0.3	0.315	0.326
		Q-Statistics	0.992	0.308	0.689	0.83	0.882	0.747
		Disagreement	0.016	0.079	0.089	0.056	0.085	0.016
		Accuracy	0.93	0.954	0.954	0.95	0.953	0.952
Homo Sapiens	Acceptor	Double fault	0.034	0.003	0.015	0.01	0.013	0.015
		Correlation	0.702	0.19	0.325	0.338	0.353	0.399
		Q-Statistics	0.989	0.555	0.667	0.978	0.844	0.978
		Disagreement	0.028	0.022	0.083	0.037	0.069	0.037
		Accuracy	0.894	0.938	0.938	0.939	0.937	0.933
	Donor	Double fault	0.022	0.001	0.008	0.007	0.01	0.008
		Correlation	0.665	0.103	0.289	0.298	0.338	0.315
		Q-Statistics	0.989	0.274	0.773	0.894	0.978	0.907
		Disagreement	0.022	0.057	0.024	0.025	0.033	0.025
		Accuracy	0.907	0.952	0.952	0.951	0.949	0.946
Average	Acceptor	Double fault	0.03	0.002	0.01	0.02	0.01	0.012
		Correlation	0.66	0.16	0.30	0.31	0.32	0.38
		Q-Statistics	0.955	0.466	0.58	0.87	0.793	0.931
		Disagreement	0.027	0.012	0.070	0.033	0.040	0.030
		Accuracy	0.830	0.941	0.940	0.940	0.940	0.930
	Donor	Double fault	0.015	0.012	0.010	0.004	0.006	0.008
		Correlation	0.599	0.09	0.260	0.240	0.28	0.28
		Q-Statistics	0.9256	0.193	0.630	0.700	0.770	0.710
		Disagreement	0.017	0.045	0.040	0.030	0.040	0.020
		Accuracy	0.920	0.954	0.950	0.950	0.950	0.950

This table depicts the five-fold Cross-validation Results, average result across the organism distribution, evaluation metrics and the ensemble combinations considered. Results highlighted in black shows the best average evaluation metrics. ENS1 consist of DNN1, DNN2, DNN3, DNN4; ENS2 consists OF CNN1, CNN2, CNN3, CNN4; ENS3 consists of DNN1, DNN2, DNN3, DNN4, CNN1, CNN2, CNN3, CNN4; ENS4 consists of CNN1, CNN2, CNN3, DNN1, DNN3; ENS5 consist of DNN1, DNN3, DNN4, CNN1, CNN2, CNN3; ENS6 includes the DNN1, DNN3, DNN4, CNN1, CNN2

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com