Identifying time-delayed gene regulatory networks via an evolvable hierarchical recurrent neural network

Table 6 The effect of network size on GRN reconstruction in case of non-linearity between the genes. Results are the average of the accuracy in terms of the Delay and Effect criterion for GRN reconstruction of 10 different randomly generated synthetic networks

	Methods	P	TP	FP	FN	Precision	Recall	F-score	Nominal p-value	Adjusted p-value
Delay	HRNN	5	4.7	7.0	5.6	0.40	0.45	0.42
		10	3.9	19.0	15.1	0.17	0.19	0.18
		20	8.0	37.8	31.6	0.17	0.20	0.18
		30	9.0	55.4	48.0	0.14	0.16	0.15
Delay	HCC-CLINDE	5	2.6	2.3	7.7	0.57	0.25	0.34	0.25	1
		10	2.6	2.3	7.7	0.57	0.25	0.34	0.69	1
		20	3.9	9.6	35.7	0.28	0.09	0.14	0.27	1
		30	3.3	13.1	53.7	0.21	0.05	0.08	0.03	0.12
Effect	HRNN	5	4.8	4.3	6.0	0.52	0.44	0.47
		10	4.6	15.0	14.5	0.23	0.23	0.23
		20	9.8	32.3	29.9	0.23	0.24	0.23
		30	9.9	51.7	47.3	0.16	0.17	0.17
Effect	HCC-CLINDE	5	2.7	2.2	7.6	0.59	0.26	0.35	0.08	0.32
		10	3.1	4.6	15.9	0.43	0.16	0.23	0.95	1
		20	3.9	9.6	35.7	0.28	0.09	0.14	0.02	0.08
		30	3.4	13.0	53.6	0.21	0.05	0.09	0.01	0.04

The p-values are for how the Delay and Effect F-scores of HCC-CLINDE method compare with HRNN. P is the number of genes for each of the networks. The variance of the noise is equal to 1

ISSN: 1756-0381