Avoiding hERG-liability in drug design via synergetic combinations of different (Q)SAR methodologies and data sources: a case study in an industrial setting

Table 4 Combining public and private data for the RF model

Public + private (RF)	ACC	BA	SENS	SPEC	PPV	NPV	MCC	KAPPA
5:0 (100% public)	0.74	0.57	0.26	0.88	0.38	0.81	0.16	0.15
5:1	0.83	0.73	0.56	0.91	0.63	0.88	0.49	0.49
5:2	0.83	0.73	0.56	0.91	0.63	0.88	0.49	0.49
5:3	0.82	0.72	0.53	0.90	0.61	0.87	0.45	0.45
5:4	0.82	0.72	0.53	0.91	0.62	0.87	0.46	0.47
5:5	0.83	0.73	0.56	0.91	0.64	0.88	0.49	0.49
4:5	0.82	0.72	0.54	0.90	0.61	0.87	0.47	0.46
3:5	0.83	0.73	0.54	0.91	0.62	0.88	0.47	0.47
2:5	0.84	0.75	0.59	0.91	0.65	0.89	0.52	0.51
1:5	0.82	0.71	0.51	0.91	0.62	0.87	0.46	0.45
0:5 (100% private)	0.82	0.73	0.57	0.89	0.61	0.88	0.48	0.47

ISSN: 1758-2946