ABSTRACT: QSAR (Quantitative Structure-Activity Relationship) analysis of monoketone curcumin analogs as prostate, pancreas, and colon anticancer agents has been conducted using electronic and molecular descriptors. This reserch aimed to obtain new monoketone curcumin analogs that had higher activity than the compounds that have been there before. Optimization molecules has been performed using the AM1 method for modeling the structure of the curcumin analogues. QSAR models derived from multiple linear regression (MLR) analysis by using a backward method. The best QSAR equation for prostate, pancreatic and colon anticancer activity is: LogIC50 = 10.385 + (1.848 x qC4) + (13.166 x qC5) - (25.067 x qC9)+(5.752 x C10) - (0.204 x µ) + (0.172 x log P) n = 33; r = 0,852; r2train = 0,727; SEE = 0,2535; FhitFtab = 4,4507 LogIC50 = 9.454 - (20.839 x qC8) + (4.855 x C10) - (7.031 x qC12) + (7.033 x O18) - (0.209xµ) + (0.108 x log P) n = 33; r = 0,873; r2train = 0,763; SEE = 0,2108; Fhit/Ftab = 5,3854 LogIC50 = 6.418 - (11.065 x qC6) - (21.153 x qC7) - (12.78 x qC9) + (11.436 x qO18) - (0.231 x µ) + (0.062 x log P) n = 33; r = 0,838; r2train = 0,703; SEE = 0,2791; Fhit/Ftab = 3,9651 The new design of curcumin analog with the best activity prediction are 4,4' - ((1E,1'E) - (4 - oxo - 2H - pyran - 3,5 (4H, 6H) - diylidene) bis (methanylylidene)) bis (2-aminobenzaldehyde); 4,4' - ((1E, 1'E) - (4 - oxo - 2H - pyran - 3,5 (4H, 6H) - diylidene) bis (methanylylidene)) bis (2, 6-dihydroxybenzaldehyde); and 4,4' - ((1E, 1'E) - (4 - oxo - 2H - pyran - 3,5 (4H, 6H) - diylidene) bis (methanylylidene)) bis (2, 6 - dimethoxybenzaldehyde).
