Table 3.

ANOVA results and predicted responce model with interaction terms for trans-resveratrol_(HPLC) amount for US treatment (Backward Elimination Regression)

	Y1:trans-resveratrol (HPLC)
	Hierarchial Terms added after Backward Elimination Regression: β 1β 3, β 1β 5, β 2β 5, β 3β 4
	SSE	df	F value	p value
Model	22.62	19	17.16	< 0.0001
x 1	1.56	1	22.52	0.0005
x 2	5.93	2	42.74	< 0.0001
x 3	1.55	2	11.15	0.0018
x 4	1.06	1	15.32	0.0021
x 5	0.016	1	0.23	0.6374
x 1 x 2	7.06	2	50.87	< 0.0001
x 1 x 3				0.1795
x 1 x 4	0.67	1	9.66	0.0090
x 1x 5				0.2952
x 2 x 3	1.62	4	5.83	0.0076
x 2 x 4	0.97	2	6.97	0.0098
x 2 x5				0.9468
x 3 x 4				0.5168
x 3 x5	3.83	2	27.58	< 0.0001
x 4 x5	0.51	1	7.33	0.0191
Residual	0.83	12
Lack of Fit	0.83	11	1.96	0.069
Pure Error	0.07	1
Cor Total	23.45	31
R2	0.9645
R2 Adj	0.9083
Adeq Precision	15.679
Significant Model Terms	X1, X2, X3, X4, X1X2, X1X4, X2X3, X2X4, X3X5, X4X5
The predicted responce model with interaction terms by D-optimal-after Backward Elimination	Ytrans-resveratrol=+0.90 − 0.23X1 + 0.59X2[1] − 0.054X2[2] − 0.35X3[1] + 0.49X3[2] + 0.44X4 + 0.084X5 − 0.76X1X2[1] + 0.061X1X2[2] − 0.15X1X4 − 0.15X2[1]X3[1] − 0.19X2[2]X3[1] + 0.55X2[1]X3[2] + 0.014X2[2] X3[2] +0.28 X2 [1]X4 + 0.10* X2 [2]X4 − 0.52 X3 [1]X5 + 0.84 X3 [2]X5 + 0.22 X4

Bold p values represents the statisticaly significant terms (p < 0.05)