Table 2.

Parameters of the best leaves of the trees for all datasets

Target	HB								Coulomb contacts
	Not filtered				Filtered				Not filtered				Filtered
	Active from all active, %		Active from all in the leaf, %		Active from all active, %		Active from all in the leaf, %		Active from all active, %		Active from all in the leaf, %		Active from all active, %		Active from all in the leaf, %
DHFR	85.1	±0.22	43.01	±0.86	77.82	±0.22	55.54	±0.30	90	±0.10	89.82	±1.12	89.36	±0.29	95.04	±0.53
Thrombin	81.52	±0.21	26.2	±0.25	78.03	±0.36	28.89	±0.39	83.64	±0.51	26.65	±0.62	80	±0.21	36.78	±1.61
Trypsin	89.56	±0.34	41.85	±0.12	93.33	±0.00	44.83	±0.27	67.33	±0.39	25.99	±0.16	68.89	±0.00	27.88	±0.21
PR	51.43	±0.77	14.54	±1.04	52.86	±0.62	14.48	±0.60	28.57	±0.00	10.26	±0.87	31.79	±2.75	10.22	±1.58
HIV protease	39.44	±1.77	9.63	±1.01	50	±0.83	19.14	±0.54	28.15	±4.24	9.79	±1.94	67.96	±1.56	18.53	±2.25
HIVRT	34.39	±2.98	10.8	±1.59	38.05	±1.62	10.14	±0.57	37.56	±4.52	14.57	±1.02	47.07	±2.90	15.46	±1.16
CDK2	32.2	±3.94	6.55	±0.80	54.6	±2.87	5.2	±0.18	18	±4.82	8.18	±4.61	20.6	±2.54	2.71	±0.28
COX1	17.69	±3.53	4.18	±0.71	50	±0.00	2.45	±0.00	27.31	±4.02	11.91	±2.83	61.54	±0.00	3.51	±0.04
HSP90	22.8	±3.15	7.37	±2.39	26	±2.29	8.73	±3.13	24	±3.65	12.34	±2.25	55.2	±3.21	9.05	±0.83
	Lipophilic contacts 1								Lipophilic contacts 2
DHFR	17.33	±1.62	23.56	±6.29	15,35	±0.89	10.81	±4.65	16.49	±2.40	11.6	±3.98	26.53	±1.01	15.47	±0.77
Thrombin	40.15	±1.77	13.51	±0.83	44.85	±0.35	13.5	±0.92	51.21	±0.40	13.05	±0.24	80.3	±0.00	6.56	±0.00
Trypsin	68.22	±1.66	30,71	±0.63	67.33	±1.80	32.03	±0.44	54.44	±2.56	21.85	±0.25	80	±0.00	5.86	±0.00
PR	24.64	±3.44	11.88	±3.73	32.86	±1.72	5.42	±2.53	26.07	±1.99	6.99	±2.03	48.93	±2.05	4.46	±0.43
HIV protease	25.56	±2.63	8.78	±1.07	19.63	±1.69	6	±0.58	41.85	±3.71	11.91	±1.06	29.63	±1.44	3.24	±0.36
HIVRT	27.32	±2.13	6.68	±1.92	38.29	±3.79	5.58	±0.41	26.34	±5.54	8.77	±1.48	35.37	±2.50	4.46	±0.44
CDK2	29	±2.24	8.87	±2.71	29.8	±3.63	4.5	±0.61	17.2	±3.49	6.17	±1.18	36	±0.00	2.11	±0.08
COX1	21.92	±1.76	6.17	±1.45	30.38	±6.07	6.13	±2.09	20	±1.54	5.88	±0.62	24.62	±1.88	4.71	±0.80
HSP90	24.8	±3.56	7.24	±1.56	20	±2.53	12.82	±4.48	20.8	±1.66	6.77	±4.29	21.2	±2.15	5.05	±2.39
	All atom contacts
DHFR	72.92	±1.37	35.08	±1.29	72.48	±0.38	44.32	±0.85
Thrombin	11.97	±1.31	7.86	±2.50	73.33	±1.46	5.71	±0.44
Trypsin	44	±4.18	34.63	±4.45	33.78	±3.10	15.45	±4.67
PR	36.07	±2.72	19.72	±2.11	31.43	±0.81	3.26	±0.19
HIV protease	21.48	±2.95	6.79	±1.17	22.78	±1.90	7.45	±0.75
HIVRT	36.34	±4.29	10.31	±0.99	53.66	±0.00	4.69	±0.04
CDK2	20	±2.37	10.77	±2.48	33.2	±2.23	4.18	±0.43
COX1	29.62	±2.46	10.57	±1.94	50.38	±2.07	5.79	±0.36
HSP90	20	±1.79	5.71	±2.59	24.4	±0.77	5.58	±5.43

Notes: Average values over ten SOM runs are calculated and RMSD values are pointed out. Data for the clustering before and after application of filters is presented. The best results are pointed out in bold for each target.

Abbreviation: HB, hydrogen bonds; SOM, self-organizing maps; RMSD, root mean square deviation.