Abstract
The present CoMFA study on thiazolidine-2,4-diones attempts to quantitatively describe the structural features of the compounds for high PPAR-γ agonistic action. A set of 50 compounds was used for the analysis. A training set and a test set were designed consisting of 39 and 11 compounds, respectively. All the molecules were modeled and energy minimized using SYBYL 6.7 software. The molecules were aligned using the most active analog as the template by field fit method. PLS (partial least square) analysis was performed on the training set and statistical parameters like q2, r2, F value and P value were calculated. A final model was generated by excluding four outliers from the training set. Cross Validation was done by leave one out method. The developed CoMFA model showed a good external as well as internal predictivity in most of the compounds with low residual values. The q2 value of the developed model was 0.549 and r2 value was 0.976. Contour plots generated from the steric and electrostatic contributions of this model were used to derive structure-activity relationships and to design a hypothetical molecular structure, which can be used as a lead for design and syntheses of newer molecules.
Keywords: CoMFA, 3D-QSAR, PPAR-γ agonists, Thiazolidine-2,4-diones, Antihyperglycemics
3