CoMFA studies require that the 3D structures of the molecules to be analyzed be aligned according to a suitable conformational
template, which is assumed to be a “bioactive” conformation. Molecular alignment was carried out using the SYBYL “fit-atom” alignment function (Tripos Inc. 2002). The crystal structure of compound 4 was used as the alignment template. Figure 1 shows the 3D alignment of 27 molecules according to the alignment scheme in Fig. 2. Fig. 1 The 3D alignment of the 27 molecules is shown by capped sticks without hydrogens Fig. 2 Molecule 4 with atoms used for superimposition NU7026 clinical trial are named 1 to 7 CoMFA study The CoMFA descriptors were used as independent variables, and pEC50 values where used as dependent variables, in partial least squares (PLS) (Wold et al., 1984) regression analysis to derive 3D QSAR models. The steric (Lennard-Jones) and electrostatic (Coulomb) CoMFA fields were calculated using an sp 3 carbon as the steric probe atom and a +1 charge for the electrostatic probe. A grid spacing of 2 Å and a distance-dependent JQ-EZ-05 dielectric constant were chosen. The cutoff value for both steric and electrostatic interactions was set to 30 kcal/mol. Partial least squares analysis PLS regression analyses were performed using cross-validation to evaluate the predictive ability of the CoMFA models. Initial
PLS regression analyses were performed in conjunction with the cross-validation (leave-one-out method) option to obtain the optimal number of components to be used in the subsequent analysis of the dataset. All the leave-one-out Luminespib research buy cross-validated PLS analyses were performed with a column filter value of 2.0 kcal/mol to improve the signal-to-noise ratio by omitting those lattice points whose energy variation was below this threshold value. The final PLS regression analysis with 10 bootstrap
groups and the optimal number of components was performed on the complete dataset. The optimal number of components was determined by selecting the smallest PRESS value. Usually this value corresponds to Unoprostone the highest cross-validated \( r^2 \left(r^2_\textcv \right) \) value. The \( r^2_\textcv \) was calculated using the formula $$ r^2_\textcv = 1 – {\frac{{\sum {} \left(Y_\textpredicted – Y_\textobserved \right)^2}}{{\sum {} \left(Y_\textobserved – Y_\textmean \right)^2}}} $$where Y predicted, Y observed, and Y mean are the predicted, actual, and mean values of the target property (pEC50), respectively. The number of components obtained from the cross-validated analysis was subsequently used to derive the final QSAR models. In addition to \( r^2_\textcv \), the corresponding PRESS [PRESS = ∑(Y predicted − Y observed)2], the number of components, the nonconventional correlation coefficient \( r^2_\textncv \), and its standard errors were also computed.