Table 3.

Amount of predicted information on gene function, measured using the ‘information accretion’ methodology and expressed as bits per gene.

Dataset	Stringency (Precision score threshold)	Method	Known annotations (bits/gene)	Recovered known annotations (bits/gene)	Newly predicted annotations (bits/gene)
Prokaryotes	0.5	10-NN	23.19	4.31	2.69
		GFP		7.71	6.89
		NFP		10.22	9.44
	0.8	10-NN		1.16	0.15
		GFP		3.09	0.77
		NFP		4.78	1.14
Fungi	0.5	10-NN	27.91	1.08	0.7
		GFP		1.96	1.37
		NFP		2.59	2
	0.8	10-NN		0.01	0.001
		GFP		0.39	0.087
		NFP		0.45	0.093
Metazoa	0.5	10-NN	19.05	1.01	0.7
		GFP		1.34	0.99
		NFP		1.56	1.25
	0.8	10-NN		0.01	0.002
		GFP		0.03	0.003
		NFP		0.08	0.014

10-NN, ten nearest neighbors; GFP, Gaussian Field Label Propagation (network-based approach); NFP, neighborhood function profile.