Discovery of novel JAK1 inhibitors through combining machine learning, structure-based pharmacophore modeling and bio-evaluation

Table 5 Performance of the different receptor-ligand pharmacophore models

No	Feature	Selectivity score	Cutoff	Precision	Recall	F1 score	Mcc
6TPF 01	DDDHH	10.828	0.000507	0.7736	0.03208	0.06161	0.1047
6TPF 02	ADDHH	9.9146	0.6751	0.8260	0.07799	0.1425	0.1802
6TPF 03	DDDH	9.3133	1.0060	0.7778	0.02556	0.04949	0.09404
6TPF 04	DDDH	9.3133	1.0049	0.7904	0.08163	0.1480	0.1742
6TPF 05	DDHH	8.3998	1.3878	0.7175	0.1067	0.1857	0.1744
6TPF 06	DDHH	8.3998	2.0498	0.8541	0.09468	0.1705	0.2079
6TPF 07	DDHH	8.3998	2.4278	0.6014	0.06573	0.1185	0.09915
6TPF 08	ADDH	8.3998	1.5159	0.8152	0.2267	0.3547	0.3131
6TPF 09	ADDH	8.3998	1.55616	0.6935	0.1328	0.2229	0.1859
6TPF 10	ADHH	7.4863	1.4717	0.6248	0.2572	0.3644	0.2251

The bold indicates the optimal model of different receptor-ligand pharmacophore models

ISSN: 1479-5876