. 2014 Jul 24;9(7):e100386. doi: 10.1371/journal.pone.0100386

Table 1. Characteristics of QSAR models.

Model Number	Descriptors	Compound’s Number	R²	Q²	Fisher	SD	Number of Variables	R² test,L30%Out	R² test,L30%Act.Dis
Model 1	QNA,V	27	0.782	0.576	12.766	0.186	6	0.524	0.604
Model 2	QNA,Lip	27	0.715	0.561	7.452	0.209	7	0.711	0.801
Model 3	QNA,Lip	27	0.711	0.548	7.363	0.21	7	0.588	0.658
Model 4	QNA,Lip	27	0.782	0.612	9.198	0.198	8	0.624	0.724
Model 5	QNA,Lip	27	0.797	0.586	11.574	0.186	7	0.543	0.623
Model 6	QNA	27	0.725	0.515	8.039	0.21	7	0.635	0.705
Model 7	QNA	27	0.696	0.528	8.504	0.215	6	0.622	0.712
Model 8	QNA	27	0.799	0.682	13.989	0.177	6	0.582	0.672
Model 9	QNA,Lip	27	0.808	0.628	12.378	0.181	7	0.677	0.747
Model 10	QNA,Lip	27	0.748	0.577	7.42	0.204	8	0.627	0.727
Model 11	QNA,Lip	27	0.809	0.628	12.383	0.181	7	0.646	0.726
Model 12	QNA	27	0.74	0.527	8.618	0.206	7	0.62	0.72
Model 13	QNA	27	0.701	0.501	6.14	0.221	8	0.827	0.907
Model 14	QNA	27	0.774	0.588	8.766	0.201	8	0.501	0.591

V – Volume descriptor, Lip – logP descriptor, L30% Out – leave-30%-out cross-validation procedure, L30%Act.Dis – leave-30%-out splitting procedure by activity distribution.