The index of ideality of correlation: A statistical yardstick for better QSAR modeling of glucokinase activators

Glucokinase is an enzyme which is responsible for the conversion of glucose to glucose-6-phosphate through ATP-dependent phosphorylation and has a significant role in glycogen synthesis and hepatic glucose production. Allosteric activators of glucokinase could be an attractive approach for the treatment of T2DM (type 2 diabetes mellitus). Recently, an innovative standard Index of Ideality of Correlation has been introduced for the estimation of QSAR (quantitative structural activity relationship) model's potential. In the present work, QSAR models for activators of glucokinase have been developed with target function TF1 and TF2 using index of ideality of correlation (IIC). Along with this, prediction of calibration sets for different QSAR models generated for different splits is also categorized as correct and wrong. Moreover, dispersion in the different runs of same split is also explained. The values of criteria R-2 and IIC for best split prepared with target function TF1 are 0.6554 and 0.7912 and that for TF2 are 0.9531 and 0.9758, respectively. The models developed with index of ideality of correlation are better than the models generated without index of ideality of correlation. The IIC could be a better criteria option for predictability of QSAR model for glucokinase activators.