Gene regulatory networks modelling using a dynamic evolutionary hybrid

BMC Bioinformatics

Table 2 Information regarding the increase of efficiency and simultaneous decrease of complexity throughout the various training phases of ENFRN

			Composite Score Values in ENFRN Structures			Number of Rules and Output Nodes in ENFRN Structures
a/a	Regulator	Target				Initial		Simplified
			Initial	Simplified	Trained	Rules	Output	Rules	Output
1	HTB1	HTA1	0.899	0.935	0.533	12	9	12	9
2	HTA1	HHF1	1.056	0.988	0.435	11	8	9	8
3	HHF1	HTA2	0.644	0.698	0.595	12	10	11	10
4	HTB1	HHF1	0.842	0.815	0.536	12	8	10	8
5	HHF1	HTB2	0.654	0.642	0.560	12	8	10	8
6	HHF2	HTA1	0.741	0.773	0.618	14	10	12	10
7	HHF2	HTA2	0.657	0.669	0.543	14	10	12	10
8	HHF2	HHF1	0.748	0.785	0.503	14	9	12	8
9	HHT1	HTA1	0.898	0.916	0.664	12	9	10	9
10	HHT1	HTB1	1.095	1.097	0.811	12	10	10	9
11	HHT1	HTB2	0.632	0.660	0.567	12	8	10	8
12	HHT1	HHF2	0.841	0.791	0.755	12	9	10	9
13	HTA1	HHT2	1.233	1.188	0.606	11	9	8	8
14	HHT2	HHF1	0.658	0.625	0.567	14	9	14	9
15	HHF2	HHT2	0.890	0.881	0.589	14	10	13	10
16	HHT2	HHT1	0.962	0.975	0.749	14	10	10	9

We can depict a decrease at the complexity levels (i.e. 87% of the 16 extracted interactions described in table) between the initial and the simplified structures followed by a corresponding decrease in the score levels between the phases of the initial and the trained ENFRN.

ISSN: 1471-2105