Journal of Theoretical & Computational Science
Open Access
ISSN: 2376-130X
ISSN: 2376-130X
Sanmati KJ, Achal M
A three dimensional quantitative structure activity relationship (3D QSAR) using k nearest neighbor molecular field analysis (kNN MFA) method was performed on a series of isatin derivatives as carboxylesterase (CE) inhibitors. This study was performed with 49 compounds (data set) using sphere exclusion (SE) algorithm for the division of the data set into training and test set. SE algorithm allows constructing training sets covering all descriptor space areas occupied by representative points. Between 3.0 to 5.5 dissimilarity levels which comprises of test set size 4 to 10, kNN-MFA methodology with stepwise (SW), simulated annealing (SA) and genetic algorithm (GA) was used for building the QSAR models. Four predictive models were generated with SW-kNN MFA (pred_r2=0.7552 to 0.9376), three predictive models were generated with SA-kNN MFA (pred_r2=0.7019 to 0.9367) and two predictive models were generated with GA-kNN MFA (pred_r2=0.8226 to 0.8497). Most significant model generated by stepwise kNNMFA showed internal predictivity 82.11% (q2=0.8211) and external predictivity 93.76% (q2=0.9376). In this model hydrophobic and steric interactions dominate the CE inhibitory activity. Hydrophobic field descriptor (H_977) with positive range indicates that positive hydrophobic potential is favorable for increase in activity and hence more hydrophobic substituent group is preferred in that region. Steric field descriptor (S_619) with negative range indicates that negative steric potential is favorable for increase in activity and hence less bulky substituent group is preferred in that region. The kNN-MFA contour plots provided further understanding of the relationship between structural features of substituted isatin derivatives and their activities which should be applicable to design newer potential CE inhibitors.