ICON-GEMs: integration of co-expression network in genome-scale metabolic models, shedding light through systems biology

Table 4 Predictive accuracy when applying various biological networks incorporating with the genome-scale metabolic model in E. coli

Conditions	DataE1 (Own)	DataE2 (Microarray)	DataE3 (RNAseq)	PPI	Random (% of edges)			Random network
Conditions	DataE1 (Own)	DataE2 (Microarray)	DataE3 (RNAseq)	PPI	25	50	75	Random network
WT 0.2/h	0.9443	0.9011	0.8852	0.8881	0.9401	0.9438	0.8869	0.8895
pgm	0.9639	0.9313	0.8064	0.7516	0.9556	0.9115	0.9215	0.9193
pgi	0.9107	0.8979	0.8888	0.8894	0.9106	0.9007	0.9105	0.9133
gapC	0.9287	0.8901	0.8741	0.8741	0.9215	0.8753	0.8753	0.8796
zwf	0.8828	0.8412	0.8252	0.8139	0.8722	0.8275	0.8319	0.8311
rpe	0.8731	0.8243	0.7918	0.8286	0.8631	0.8071	0.8071	0.8189
WT 0.5/h	0.9656	0.8660	0.9001	0.8660	0.9588	0.9005	0.9007	0.9025
WT 0.7/h	0.8962	0.8357	0.8643	0.8357	0.9088	0.8410	0.8684	0.8662
Mean	0.9206	0.8434	0.8545	0.8434	0.9164	0.8759	0.8753	0.8775
Standard deviation	0.0333	0.0354	0.0383	0.0435	0.0332	0.0439	0.0366	0.0345

Eight distinct conditions encompass wild-type E. coli growth rates at 0.2 (as reference (RF)), 0.5 (WT0.5), and 0.7 (WT0.7) per hour as well as specific gene deletions (of genes pgm, pgi, gapc, zwf, and rpe)

ISSN: 1471-2105