. 2023 Sep 25;28(19):6793. doi: 10.3390/molecules28196793

Table 1.

Performance comparison of IonPred with other tools on metal-ion test sets.

Ion	Method	Rec	Pre	F1	MCC	AUC	AUPR
	MIB	0.739	0.220	0.339	0.389	0.922	0.388
	TargetS	0.450	0.750	0.563	0.578	0.868	0.594
Zn²⁺	ZinCaps	0.753	0.780	0.766	0.601	0.915	0.768
	IonCom	0.779	0.137	0.233	0.317	0.892	0.671
	LMetalSite	0.722	0.859	0.785	0.760	0.971	0.801
	IonPred	0.790	0.840	0.814	0.600	0.958	0.780
	MIB	0.341	0.082	0.132	0.139	0.764	0.105
	TargetS	0.119	0.487	0.191	0.244	0.775	0.165
Ca²⁺	DELIA	0.172	0.630	0.270	0.330	0.782	0.251
	IonCom	0.297	0.247	0.270	0.258	0.697	0.166
	LMetalSite	0.413	0.720	0.525	0.540	0.904	0.490
	IonPred	0.467	0.759	0.578	0.615	0.923	0.520
	MIB	0.246	0.043	0.073	0.082	0.673	0.053
	TargetS	0.118	0.491	0.190	0.237	0.715	0.148
Mg²⁺	IonCom	0.240	0.250	0.245	0.237	0.685	0.184
	DELIA	0.129	0.065	0.086	0.287	0.740	0.198
	LMetalSite	0.245	0.728	0.367	0.419	0.866	0.316
	IonPred	0.400	0.780	0.529	0.470	0.889	0.450
	MIB	0.462	0.096	0.159	0.193	0.855	0.168
	TargetS	0.271	0.496	0.350	0.362	0.862	0.322
Mn²⁺	DELIA	0.502	0.665	0.572	0.574	0.887	0.489
	IonCom	0.511	0.245	0.331	0.344	0.831	0.304
	LMetalSite	0.613	0.719	0.662	0.661	0.963	0.625
	IonPred	0.620	0.700	0.658	0.670	0.970	0.670
	MIB	0.586	0.620	0.603	0.573	0.909	0.354
Fe²⁺	TargetS	0.345	0.254	0.293	0.245	0.760	0.299
	IonPred	0.749	0.728	0.738	0.723	0.937	0.771
	IonCom	0.610	0.498	0.548	0.579	0.909	0.567
Fe³⁺	MIB	0.474	0.399	0.433	0.383	0.813	0.438
	IonPred	0.743	0.612	0.671	0.652	0.928	0.724
	IonCom	0.596	0.398	0.477	0.592	0.890	0.399
Cu²⁺	MIB	0.466	0.280	0.350	0.358	0.870	0.419
	IonPred	0.789	0.634	0.703	0.620	0.939	0.677
	IonCom	0.210	0.178	0.193	0.160	0.723	0.156
K⁺	TargetS	0.389	0.411	0.400	0.341	0.876	0.336
	IonPred	0.498	0.672	0.572	0.524	0.912	0.478
Na⁺	IonCom	0.451	0.292	0.355	0.218	0.709	0.233
	IonPred	0.523	0.731	0.610	0.595	0.904	0.487

Rec refers to Recall, Pre refers to Precision, MCC refers to Matthew’s correlation coefficient, F1 refers to F1 score, AUC refers to Area under the curve, and AUPR refers to Area under precision recall curve. Bold font indicates metric with best performance.