Advancing drug–target interaction prediction: a comprehensive graph-based approach integrating knowledge graph embedding and ProtBert pretraining

Djeddi, Warith Eddine; Hermi, Khalil; Ben Yahia, Sadok; Diallo, Gayo

doi:10.1186/s12859-023-05593-6

Table 9 Performance comparison of DTIOG variants on the GPCR dataset using ProtBERT

From: Advancing drug–target interaction prediction: a comprehensive graph-based approach integrating knowledge graph embedding and ProtBert pretraining

Approaches	AUPR	AUC	ACC	MCC	F1
Cosine similarity
DTIOG-RF	0.930	0.969	0.956	0.899	0.925
DTIOG-DT	0.948	0.978	0.968	0.926	0.945
DTIOG-MLP	0.836	0.936	0.874	0.673	0.755
DTIOG-KNeighbors	0.946	0.977	0.939	0.865	0.899
DTIOG-BaggingClassifier	0.635	0.500	0.683	0.000	0.042
DTIOG-GradientBoosting	0.792	0.923	0.872	0.669	0.755
DTIOG-GaussianNB	0.431	0.628	0.439	0.108	0.437
DTIOG-SGD	0.454	0.691	0.720	0.274	0.443
DTIOG-ETC	0.951	0.980	0.971	0.931	0.949
Euclidean distance
DTIOG-RF	0.950	0.979	0.970	0.930	0.948
DTIOG-DT	0.963	0.985	0.978	0.948	0.961
DTIOG-MLP	0.995	0.998	0.993	0.983	0.987
DTIOG-KNeighbors	0.961	0.984	0.959	0.905	0.930
DTIOG-BaggingClassifier	0.635	0.500	0.729	0.000	0.000
DTIOG-GradientBoosting	0.894	0.973	0.936	0.842	0.885
DTIOG-GaussianNB	0.597	0.772	0.752	0.436	0.604
DTIOG-SGD	0.594	0.805	0.784	0.437	0.572
DTIOG-ETC	0.957	0.982	0.974	0.940	0.955
Manhattan distance
DTIOG-RF	0.950	0.979	0.970	0.930	0.948
DTIOG-DT	0.959	0.983	0.976	0.943	0.958
DTIOG-MLP	0.994	0.998	0.993	0.983	0.987
DTIOG-KNeighbors	0.959	0.983	0.953	0.894	0.921
DTIOG-BaggingClassifier	0.635	0.500	0.683	0.000	0.042
DTIOG-GradientBoosting	0.872	0.966	0.931	0.828	0.875
DTIOG-GaussianNB	0.586	0.775	0.755	0.436	0.602
DTIOG-SGD	0.591	0.803	0.781	0.426	0.568
DTIOG-ETC	0.960	0.983	0.976	0.944	0.958
Jaccard similarity
DTIOG-RF	0.938	0.974	0.962	0.912	0.934
DTIOG-DT	0.958	0.983	0.975	0.941	0.956
DTIOG-MLP	0.984	0.996	0.989	0.974	0.981
DTIO-KNeighbors	0.968	0.987	0.965	0.918	0.939
DTIOGBaggingClassifier	0.635	0.500	0.729	0.000	0.000
DTIO-GradientBoosting	0.880	0.969	0.923	0.808	0.861
DTIO-GaussianNB	0.473	0.660	0.471	0.138	0.448
DTIO-SGD	0.560	0.780	0.775	0.400	0.536
DTIOG-ETC	0.954	0.981	0.972	0.934	0.951
Pearson correlation coefficient
DTIOG-RF	0.940	0.975	0.963	0.915	0.937
DTIOG-DT	0.956	0.982	0.974	0.938	0.954
DTIOG-MLP	0.987	0.996	0.990	0.975	0.981
DTIOG-KNeighbors	0.972	0.989	0.961	0.910	0.933
DTIOG-BaggingClassifier	0.635	0.500	0.729	0.000	0.000
DTIOG-GradientBoosting	0.898	0.975	0.934	0.838	0.882
DTIOG-GaussianNB	0.497	0.676	0.527	0.189	0.468
DTIOG-SGD	0.587	0.797	0.780	0.419	0.554
DTIOG-ETC	0.958	0.983	0.975	0.940	0.956

The best results are indicated in bold concerning each row

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com

BMC Bioinformatics

Contact us