Skip to main content

Table 2 Ensembles of feasible vectors generated with different schemes

From: Predicting network modules of cell cycle regulators using relative protein abundance statistics

Ensemble

Scheme

Ensemble

Selection of

Feasibility criteria

# generations

#

S

#

#

size

the initial DE

used in selection

per DE run

DE runs

(Range of

   

population from

step of DE

  

predictions)

   

Ensemble 1

    

1

 

3146

-

Parameter vectors

Ensemble generated

-

30

    

satisfy F C 1

in optimization [22]

  

2

1

243

-

Parameter vectors

Ensemble extracted

-

51

    

satisfy F C 1

from 50,000

  
     

LHS samples

  

3

2

7143

Randomly selected

F C 1

400

1

6

   

parameter vectors

    

4

3

1893

V max (10)

F C 1

400

1

41

5

4

1594

V max (10)

F C 1 and F C 2

1600

1

64

6

4

1326

V max (10)

F C 1 and F C 2

1600

1

69

7

5

3405

V max (123)

F C 1 and F C 2

1600

1

94

8

5

3753

V max (123)

F C 1 and F C 2

1600

1

80

9

6

2207

S max & V max (123)

F C 1 and F C 2

1600

1

117

10

6

1842

S max & V max (123)

F C 1 and F C 2

1600

1

95

11

7

3704

S max & V max (123)

F C 1, F C 2, and F C 3

1600

1

112

12

7

3481

S max & V max (123)

F C 1, F C 2, and F C 3

1600

1

133

13

8

4280

S max & V max (123)

F C 1 and F C 3

1600

1

313

14

8

4550

S max & V max (123)

F C 1 and F C 3

1600

1

367

15

7

15520

S max & V max (123)

F C 1, F C 2, and F C 3

2200

4

293

16

8

15050

S max & V max (123)

F C 1 and F C 3

2200

4

671

  1. Parameter ranges used for LHS are from Ensemble 1. Parameter vectors in all ensembles capture the phenotypes listed in Additional file 1: Table S7, while missing the phenotypes in Additional file 1: Table S8. S: The range of the phenotypic prediction vectors generated per ensemble (unique rows of the prediction matrix P). V max (10): Biased selection is used to expand the estimated volume spanned by the initial population with respect to the axes of the ten most critical parameters (Table 3). V max (123): Biased selection is used to expand the estimated volume spanned by the initial population with respect to the axes of 123 kinetic parameters. S max : Biased selection is used to enhance the initial population’s range of phenotypic predictions. The prediction ranges for all ensembles can be reproduced using Additional file 4 (simulation code), and Additional files 5, 6, 7, 8 and 9 (Ensembles 1 through 16)