Improving protein function prediction using domain and protein complexes in PPI networks

BMC Systems Biology

Table 1 The results of DCS, DSCP and DSCP using protein complexes predicted by DPClus on DIP PPI

Methods	Size	MF					CC					BP
		#GO	PPV	TPR	F-measure	MCC	#GO	PPV	TPR	F-measure	MCC	#GO	PPV	TPR	F-measure	MCC
DCS	[10–30]	75	0.44	0.37	0.40	0.40	89	0.40	0.39	0.39	0.38	216	0.34	0.35	0.34	0.35
	(30–50]	10	0.41	0.36	0.39	0.38	18	0.30	0.25	0.27	0.26	36	0.31	0.30	0.30	0.30
	(50–100]	8	0.50	0.42	0.46	0.45	12	0.33	0.30	0.32	0.30	13	0.34	0.33	0.33	0.33
	(100–200]	2	0.71	0.62	0.66	0.65	5	0.27	0.24	0.25	0.22	2	0.47	0.46	0.47	0.46
	In total	95	0.45	0.38	0.41	0.41	124	0.37	0.35	0.36	0.35	267	0.34	0.34	0.34	0.34
DSCP	[10–30]	75	0.44	0.36	0.40	0.39	89	0.49	0.48	0.48	0.48	216	0.38	0.38	0.38	0.37
	(30–50]	10	0.37	0.31	0.34	0.33	18	0.36	0.33	0.34	0.33	36	0.37	0.35	0.36	0.35
	(50–100]	8	0.50	0.38	0.43	0.42	12	0.35	0.31	0.32	0.31	13	0.40	0.40	0.40	0.39
	(100–200]	2	0.73	0.55	0.63	0.62	5	0.30	0.26	0.28	0.25	2	0.55	0.53	0.54	0.53
	In total	95	0.44	0.36	0.40	0.39	124	0.45	0.43	0.44	0.43	267	0.38	0.37	0.38	0.37
DSCP_DPClus	[10–30]	75	0.44	0.36	0.40	0.39	89	0.41	0.34	0.37	0.36	216	0.34	0.32	0.33	0.33
	(30–50]	10	0.37	0.31	0.34	0.33	18	0.30	0.25	0.27	0.26	36	0.30	0.26	0.28	0.27
	(50–100]	8	0.50	0.38	0.43	0.42	12	0.32	0.30	0.31	0.29	13	0.35	0.30	0.33	0.31
	(100–200]	2	0.73	0.55	0.63	0.62	5	0.26	0.25	0.26	0.23	2	0.49	0.48	0.48	0.47
	In total	95	0.44	0.36	0.40	0.39	124	0.38	0.32	0.34	0.34	267	0.34	0.32	0.33	0.32

The GO terms in MF, CC and BP are divided into 4 groups according to the number of proteins annotated by them, respectively. Those GO terms which are consisted of more than 10 or less than 200 proteins are filtered out because they are either too rare or too general for prediction. This table shows prediction results of DCS, DSCP and DSCP using protein complexes predicted by DPClus (DSCP_DPClus) based on DIP PPI data, including the number of GO terms in each group and each category, and the corresponding values of PPV, TPR and MCC. Here, column "size" means the number of protein in each GO term. "#GO" denotes the number of GO terms.

ISSN: 1752-0509