A quantitative structure-activity relationship (QSAR) study of twenty flavonoid derivatives with antimutagenic activity against 3-nitrofluoranthene (3-NFA) was performed by Partial Least Squares (PLS), using Ordered Predictors Selection (OPS) algorithm for variable selection. Four descriptors (PJI2, Mor27m, G1e and R4u+) were selected and a good model (n = 19; R(2) = 0.747; SEC = 0.332; PRESS(cal) = 1.768; F((2,27)) = 23.585; Q(LOO)(2) = 0.590; SEV = 0.388; PRESS(val) = 2.858; R(pred)(2) = 0.591; SEP = 0.394; ARE(pred) = 5.230%; k = 1.005; k' = 0.990; |R(02) - R'(02)| = 0.109) was built with two latent variables describing 83.410% of the original information. Leave-N-out cross validation (LNO) and y-randomization were performed in order to confirm the robustness of the model. The topological descriptors selected indicate that the antimutagenic activity against 3-NFA depends on molecular size, shape and Sanderson electronegativity of flavonoids. The proposed model may provide a better understanding of the antimutagenic activity of flavonoids and can be used as a guidance for proposition of new chemopreventive agents.Copyright (c) 2010 Elsevier Masson SAS. All rights reserved.
