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

BMC Bioinformatics

Table 2 Selected EMs and the respective variance (ϑ) values for one to 10 number of n _Facs obtained for the simulated data with 10 % Gaussian noise

n _Fac	EM/ϑ	1	2	3	4	5	6	7	8	9	10
1	EM	70
1	ϑ	29.61
2	EMs	70	7
2	ϑ	29.61	54.09
3	EMs	70	7	40
3	ϑ	29.61	54.09	71.50
4	EMs	7	69	13	40
4	ϑ	24.60	47.20	65.78	83.55
5	EMs	7	71	33	13	37
5	ϑ	24.60	46.34	64.74	82.59	94.05
6	EMs	7	33	13	3	37	23
6	ϑ	24.60	44.17	62.35	79.05	91.47	97.09
7	EMs	7	33	13	3	37	23	12
7	ϑ	24.60	44.17	62.35	79.05	91.47	97.09	97.26
8	EMs	7	33	13	3	37	23	12	19
8	ϑ	24.60	44.17	62.35	79.05	91.47	97.09	97.26	97.30
9	EMs	7	33	13	3	37	23	12	19	16
9	ϑ	24.60	44.17	62.35	79.05	91.47	97.09	97.26	97.30	97.31
10	EMs	7	33	13	3	37	23	12	19	16	17
10	ϑ	24.60	44.17	62.35	79.05	91.47	97.09	97.26	97.30	97.31	97.32
PCA	n_lv**	1	2	3	4	5	6	7	8	9
PCA	ϑ	50.04	82.08	91.83	97.05	99.39	99.96	100.00	100.00	100.00

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]. n _lv ** number of latent variables for PCA

ISSN: 1471-2105