Table 1 Sparseness of networks and corresponding conditional FDR, z-value threshold, the number of GG, EG, EE links (n), and the range of correlations and partial correlations

From: Studying alternative splicing regulatory networks through partial correlation analysis

	Human			Mouse			Rat
Sparseness	0.02%	0.01%	0.005%	0.02%	0.01%	0.005%	0.02%	0.01%	0.005%
GG
cFDR	0.227	0.206	0.185	0.05	0.031	0.02	0.045	0.025	0.015
z	4.42	4.61	4.79	4.77	5.02	5.25	4.8	5.07	5.31
n	13,552	6,523	3,202	12,014	5,878	2,893	2,672	1,307	653
\|r_g1g2\|	≥ 0.947	≥ 0.957	≥ 0.965	≥ 0.964	≥ 0.972	≥ 0.979	≥ 0.965	≥ 0.974	≥ 0.981
EG
cFDR	0.249	0.183	0.132	0.112	0.066	0.037	0.094	0.051	0.026
z	4.27	4.52	4.76	4.5	4.78	5.05	4.55	4.85	5.13
n	264,615	123,211	57,584	246,027	117,761	57,227	50,828	23,960	11,822
\|r_e1g2\| \|r_e1g2·g1\|	≥ 0.836	≥ 0.862	≥ 0.884	≥ 0.847	≥ 0.874	≥ 0.896	≥ 0.848	≥ 0.876	≥ 0.899
EE
cFDR	0.091	0.054	0.031	0.056	0.026	0.011	0.05	0.021	0.008
z	4.49	4.77	5.03	4.63	4.95	5.26	4.65	4.98	5.32
n	1,028,385	489,485	242,567	1,110,763	535,536	263,301	215,750	106,617	52,114
\|r_e1e2\| \|r_e1e2·g1\| \|r_e1e2·g2\| \|r_e1e2·g1g2\|	≥ 0.720	≥ 0.757	≥ 0.788	≥ 0.699	≥ 0.741	≥ 0.778	≥ 0.690	≥ 0.733	≥ 0.773

The sparseness is the percentage of true links among all possible node pairs. Note that the number of declared links is very sensitive to the conditional FDR threshold. For example, for the GG network of human, when the conditional FDR (cFDR) changes from 4.42 to 4.61 (a 4% increase), the number of GG links changes from 13,552 to 6,523 (a 48% decrease). Meanwhile, the sparseness is from 0.02% to 0.01% (a 50% decrease).

Back to article page

ISSN: 1474-760X

Contact us

Submission enquiries: editorial@genomebiology.com
General enquiries: info@biomedcentral.com