An MCASE approach to the search of a cure for Parkinson's Disease.

BMC Pharmacology

Table 1 Statistics of the MAO, DA and NMDA databases

DA database	249 molecules, 135 are active, 105 are inactive and 9 are marginal
Test Sets	Sensitivity (%)^a	Specificity (%)^b	OCP (%)^c	Coverage (%)^d
1	100.00	100.00	100.00	100.00
2	86.70	88.90	87.50	100.00
3	100.00	80.00	91.67	96.00
Average	95.57	89.63	93.06	98.67
NMDA database	374 molecules, 145 are active, 211 are inactive and 18 are marginal
Test Sets	Sensitivity (%)	Specificity (%)	OCP (%)	Coverage (%)
1	85.71	88.24	87.10	88.57
2	76.92	88.24	83.33	81.08
3	81.25	93.33	87.10	88.57
Average	81.30	89.93	85.84	86.07
MAO database Selectivity approach	239 molecules, 73 are active, 124 are inactive and 2 are marginal
Test Sets	Sensitivity (%)	Specificity (%)	OCP (%)	Coverage (%)
1	100.00	100.00	100.00	100.00
2	85.70	100.00	95.00	95.24
3	100.00	100.00	100.00	91.30
Average	95.23	100.00	98.33	95.51
MAO database Activity approach	237 molecules, 98 are active, 118 are inactive and 21 are marginal
Test Sets	Sensitivity (%)	Specificity (%)	OCP (%)	Coverage (%)
1	90.00	100.00	95.65	100.00
2	91.70	100.00	95.83	100.00
3	91.70	100.00	95.45	95.65
Average	91.13	100.00	95.64	98.55

^a Ability to predict active compounds correctly ^b Ability to predict inactive compounds correctly ^c The percentage of Observed Corrected Predictions (OCP) ^d Probability for an unknown compound to receive a statistically satisfactory prediction.