Skip to main content

Table 2 Comparison of computed and estimated coefficients

From: Priming nonlinear searches for pathway identification

 

Computed coefficients

Estimated coefficients

a10

0

0.0000

a11

-14.6780

-14.3647

a12

0

-0.1466

a13

7.3390

7.3414

a14

0

-0.2165

a15

-7.3390

-7.1723

a20

0

0.0000

a21

14.6780

14.6119

a22

-14.6780

-14.6540

a23

0

-0.0009

a24

0

0.0494

a25

0

-0.0309

a30

0

0.0000

a31

0

-2.3527

a32

0

1.3989

a33

-27.2517

-27.9204

a34

0

1.7491

a35

0

-0.9955

a40

0

0.0000

a41

0

2.0843

a42

0

-1.0925

a43

18.5664

19.0295

a44

-18.5664

-20.2112

a45

-9.2832

-8.3594

a50

0

0.0000

a51

0

-0.4026

a52

0

0.1384

a53

0

-0.0059

a54

18.5664

18.8987

a55

-18.5664

-18.7852

  1. Regression coefficients for the small gene network (Figure 1), linearized about the steady state and based on relative deviations (option II). The first and second columns contain the computed and estimated regression coefficients, respectively. The regression coefficients a ij refer to the influence of variable j on variable i, while ai 0is the constant term in each regression model. As the table indicates, the correspondence is good, except for the coefficients relating to X3 and X4 (see Text for explanation). The dataset consisted of 401 data points in the interval [0,4] and resulted from a simulation in which X3 was perturbed at t = 0 to a value 5% above its steady-state value.