Predictive Modeling of the Human Hepatoma (Huh-7D12) Cancer Line of a Series of bis- (5-arylidene-rhodanine-3-yl) Diamine

This work deals with the prediction of the antiproliferative activity of eighteen (18) substances derived from bis-5-arylidene rhodanine against human hepatoma tumor line (Huh-7D12). By applying the functional density theory (DFT) method to the B3LYP / 6-31G (d, p) level, theoretical descriptors were determined and correlated with antiproliferative (Huh-7) activity by linear regression multiple (RML). This correlation has shown that the electron energy, the energy of the lowest vacant molecular orbital (ELUMO) and the molecular volume (VM) are the quantum and geometric descriptors that best influences the antiproliferative activity of the molecules studied. The coefficient of determination R2 indicates that 97.9% of the molecular descriptors defining this model are taken into account with a standard deviation of 0.015. The significance of the model reflected by the Fischer test is estimated at 123.648. The robustness of the model given by the cross-validation correlation coefficient (Q2CV) is 97.9%. This model has been validated by Tropsha criteria. The very good correlation between these three descriptors and the Huh-7 activity was confirmed by the nonlinear multiple regression (RNML) method with better statistical data. (R2 = 0,998 ; Q2CV = 0,998 ; RMSE = 0,006).