Using PPI network autocorrelation in hierarchical multi-label classification trees for gene function prediction

BMC Bioinformatics

Table 7 Basic properties of the yeast PPI networks

Dataset	% of connected genes				% of function-relevant interactions				Avg. degree of node
	FUN		GO		FUN		GO		FUN		GO
	DIP	BioGRID	DIP	BioGRID	DIP	BioGRID	DIP	BioGRID	DIP	BioGRID	DIP	BioGRID
seq	46	96	46	97	8	8	15	8	7.09	7.09	7.15	54.97
pheno	46	98	46	99	6	11	16	11	3.53	27.67	17.57	27.75
struc	13	98	59	98	7	14	14	14	7.27	54.74	7.07	54.97
hom	45	97	48	14	7	16	14	16	7.22	54.301	7.79	58.57
cellcycle	72	99	47	99	2	17	17	16	7.36	55.63	7.38	55.72
church	46	99	46	99	15	16	13	15	7.35	56.21	7.39	56.28
derisi	72	100	73	100	7	17	11	16	11.17	84.43	11.19	84.64
eisen	35	65	35	65	9	19	19	17	4.68	32.47	4.69	32.52
gasch1	47	99	47	99	9	17	19	16	7.41	55.83	7.42	55.92
gasch2	47	98	47	99	7	17	17	16	7.35	55.62	7.39	55.93
spo	48	99	48	99	3	13	17	16	7.31	55.27	7.32	55.35
exp	46	99	46	99	15	16	39	15	7.35	56.16	7.36	56.3

The percentage of connected genes, the percentage of function-relevant interactions and the average degree of nodes for 2 different PPI networks (DIP [28] and BioGRID [27].

ISSN: 1471-2105