Three-Dimensional Quantitative Structure−Activity Relationship of Interleukin 1-β Converting Enzyme Inhibitors:  A Comparative Molecular Field Analysis Study

Abstract

A three-dimensional quantitative structure-activity relationship (QSAR) study using the comparative molecular field analysis (CoMFA) method was performed on a series of interleukin 1-beta converting enzyme (ICE) inhibitors. The compounds studied have been reported to be time-dependent inhibitors of ICE. This study was performed using 49 compounds, in which the CoMFA models were developed using a training set of 39 compounds. All the compounds were modeled using the X-ray crystal structure of tetrapeptide aldehyde inhibitor/ICE complex. The inhibitor compounds were considered both as neutral species and as P1 carboxylate ionized species. Superimpositions were performed using two alignment rules, namely, an alignment of the structures based on RMS fitting of the backbone heavy atoms of each structure to compound 2 and an alignment based on SYBYL QSAR rigid body field fit of the steric and electrostatic fields of the molecules to the fields of compound 2. Use of LUMO energies or ClogP as additional descriptors in the QSAR table did not improve the significance of the CoMFA models. Steric and electrostatic fields of the inhibitors were found to be the relevant descriptors for structure-activity relationships. The predictive ability of the CoMFA model was evaluated by using a test set of 10 compounds (r2pred as high as 0.859). Further comparison of the coefficient contour maps with the steric and electrostatic properties of the receptor show a high level of compatibility.

On Selection of Training and Test Sets for the Development of Predictive QSAR models

“…PLS is normally used in combination with cross-validation to obtain the optimum number of components. This ensures that the QSAR equations are selected based on their ability to predict the data rather than to fit the data [33]. From the factor loading table (rotated component matrix), variables with high loading (> 0.7) in such factors where the binding affinity shows high or moderate loading were selected for the PLS regression.…”

QSTR with Extended Topochemical Atom (ETA) Indices 8.a QSAR for the inhibition of substituted phenols on germination rate of Cucumis sativus using chemometric tools

“…Table 7 shows that the best statistical quality along with prediction statistics with ETA descriptors is found in case of GFA-MLR equation. (1, 5, 9, 13, 17, 21, 25, 29, 33, 7, 41), (2, 6, 10, 14, ..., 34, 38), ..., (4,8,12,16,20,24,28,32,36,40). …”

“…PLS is normally used in combination with cross-validation to obtain the optimum number of components. This ensures that the QSAR equations are selected based on their ability to predict the data rather than to fit the data [40]. In case of FA-PLS analysis on the present data set, factor loading table obtained from factor analysis was used for primary variable screening.…”

Automatic generation of alignments for 3D QSAR analyses

“…The use of the routine has been reported by several workers (see, e.g., [33][34][35]) but is quite complex in operation [36], involving the inclusion of weighted field-fit energy penalties as additional parameters in the Tripos force field. More recently, Paretti et al have described a procedure that is analogous to that reported here [20] but that uses Monte Carlo and simplex procedures for the generation of the alignments, rather than a GA, and PLS analysis of N×N similarity matrices, rather than CoMFA and CoMSIA.…”