. Author manuscript; available in PMC: 2009 Oct 21.

Published in final edited form as: J Med Chem. 2008 Dec 25;51(24):7827–7833. doi: 10.1021/jm8009074

Table 3.

Results for 10 Different Systems Showing the Correlations between Enzyme and Cell Activities and the Correlations between Experimental and Predicted Cell Activities

target	cell	R², enzyme vs cell	R², enzyme test set^a	R², training set^b	R², test set^a	R², scrambled^c	descriptors tested	number of compounds
FPPS	Plasmodium falciparum	0.06	0.01	0.77	0.74	0.16	150	26
FPPS	Dictyostelium discoideum	0.49	0.46	0.70	0.70	0.07	123	102
FPPS	Leishmania donovani	0.55	0.47	0.87	0.80	0.11	138	21
UPPS	Streptococcus pneumoniae	0.03	0.10	0.71	0.69	0.14	109	27
MurI	Streptococcus pneumoniae	0.61	0.57	0.72	0.68	0.13	123	42
HCV NS3 Protease	Hu-7	0.03	0.02	0.78	0.77	0.12	142	34
HIV-1 integrase	MT4	0.12	0.04	0.69	0.64	0.13	126	30
HLG Phosphorylase	rat hepatocytes	0.55	0.48	0.71	0.74	0.19	132	35
KDR kinase	NIH 3T3	0.36	0.30	0.60	0.58	0.09	117	42
Akt Kinase	MiaPaCa-2 (pancreatic cancer)	0.41	0.35	0.68	0.68	0.12	80	48
average		0.32	0.28	0.72	0.70	0.13	124	42

Correlation obtained from a leave-two-out analysis.

Correlations based on cell activity predictions using a combinatorial descriptor search.

Average correlation obtained from a leave-two-out analysis of 10 sets of scrambled data.