An end-to-end method for predicting compound-protein interactions based on simplified homogeneous graph convolutional network and pre-trained language model

Table 6 Classification results of SPVec-SGCNs model compared with four state-of-the-art models on BindingDB and PubChem test sets

Testset	Methods	Accuracy	Precision	Recall	F1-Score	AUC	AUPR
BindingDB	Ours	0.9805	0.9763	0.9847	0.9805	0.9979	0.9979
	PMFCPI	0.8218	0.8189	0.8501	0.8234	0.8962	0.9044
	GraphCPI	0.7237	0.7478	0.7016	0.7223	0.7697	0.7734
	STCPI	0.8234	0.7965	0.7999	0.8228	0.8752	0.8745
	GcForest	0.862	0.8523	0.8547	0.8678	0.8956	0.8957
	CCL-DTI	0.8749	0.8594	0.8782	0.8631	0.9021	0.8954
	SgCPI	0.8334	0.8348	0.8329	0.8335	0.8521	0.8545
PubChem	Ours	0.9948	0.7754	0.8262	0.8000	0.9875	0.8709
	PMFCPI	0.6967	0.1893	0.6743	0.4461	0.7880	0.2243
	GraphCPI	0.6253	0.0587	0.6227	0.1048	0.7653	0.2540
	STCPI	0.8439	0.6261	0.7359	0.6007	0.8949	0.7244
	GcForest	0.6003	0.0202	0.6411	0.0391	0.6374	0.2037
	CCL-DTI	0.6482	0.1467	0.5649	0.3732	0.8036	0.1734
	SgCPI	0.8679	0.0573	0.6557	0.1064	0.8278	0.0693

ISSN: 1758-2946