Large scale comparison of QSAR and conformal prediction methods and their applications in drug discovery

Table 4 Comparison of the results obtained for the internal and the temporal validation for the QSAR and the MCP models built on ChEMBL_23, considering the 296 protein targets shared by ChEMBL_23 and ChEMBL_24

Method	Prediction set	Model targets	Confidence level (%)	Sensitivity	Specificity	CCR
QSAR	ChEMBL_23	296		0.78 (± 0.15)	0.84 (± 0.14)	0.81 (± 0.07)
QSAR	ChEMBL_24	296		0.61	0.80	0.71
MCP_incl	ChEMBL_23	296	70	0.73 (± 0.03)	0.73 (± 0.03)	0.73 (± 0.02)
MCP_incl	ChEMBL_23	296	80	0.84 (± 0.02)	0.84 (± 0.03)	0.84 (± 0.02)
MCP_incl	ChEMBL_23	296	90	0.94 (± 0.02)	0.93 (± 0.02)	0.93 (± 0.02)
MCP_incl	ChEMBL_24	296	70	0.41	0.45	0.43
MCP_incl	ChEMBL_24	296	80	0.63	0.67	0.65
MCP_incl	ChEMBL_24	296	90	0.85	0.87	0.86
MCP_excl	ChEMBL_23	296	70	0.72 (± 0.04)	0.73 (± 0.03)	0.73 (± 0.02)
MCP_excl	ChEMBL_23	296	80	0.77 (± 0.11)	0.77 (± 0.11)	0.77 (± 0.11)
MCP_excl	ChEMBL_23	296	90	0.65 (± 0.19)	0.63 (± 0.20)	0.64 (± 0.19)
MCP_excl	ChEMBL_24	296	70	0.41	0.44	0.42
MCP_excl	ChEMBL_24	296	80	0.54	0.60	0.52
MCP_excl	ChEMBL_24	296	90	0.42	0.44	0.43

ISSN: 1758-2946