Skip to main content

Table 1 Selected EMs and the respective captured variance (ϑ) values for one to 10 number of n Fac s obtained for the simulated data without noise

From: A principal components method constrained by elementary flux modes: analysis of flux data sets

n Fac EM/ϑ 1 2 3 4 5 6 7 8 9 10
1 EM 70          
1 ϑ 29.45          
2 EMs 70 7         
2 ϑ 29.45 53.82         
3 EMs 70 7 40        
3 ϑ 29.45 53.82 71.16        
4 EMs 7 69 13 40       
4 ϑ 24.49 46.86 65.83 83.55       
5 EMs 7 71 13 33 37      
5 ϑ 24.49 45.99 64.42 82.48 93.96      
6 EMs 7 33 13 3 37 23     
6 ϑ 24.49 44.02 62.29 78.93 91.37 97.08     
7 EMs 7 33 13 3 37 23 12    
7 ϑ 24.49 44.02 62.29 78.93 91.37 97.08 97.23    
8 EMs 7 33 13 3 37 23 12 19   
8 ϑ 24.49 44.02 62.29 78.93 91.37 97.08 97.23 97.26   
9 EMs 7 33 13 3 37 23 12 19 16  
9 ϑ 24.49 44.02 62.29 78.93 91.37 97.08 97.23 97.26 97.28  
10 EMs 7 33 13 3 37 23 12 19 16 17
10 ϑ 24.49 44.02 62.29 78.93 91.37 97.08 97.23 97.26 97.28 97.28
PCA nlv** 1 2 3 4 5 6 7 8 9  
PCA ϑ 50.16 82.19 91.85 97.00 99.27 99.96 100.00 100.00 100.00  
BF* EMs 70 7 40 13 37 23 14 12 8 16
BF* ϑ 29.45 53.82 71.16 82.44 90.54 91.84 92.04 92.12 92.17 92.20
  1. The set of truly active EMs for data generation was EMs = [1, 3, 7, 12, 13, 14, 16, 19, 20, 22, 23, 24, 28, 32, 33, 37]. BF* best-first identification by the greedy approach. n lv ** number of latent variables for PCA