Identifying protein complexes based on node embeddings obtained from protein-protein interaction networks

BMC Bioinformatics

Table 1 Performance comparison results on HPRD and DIP datasets

Methods	No. of complexes	Precision	Recall	F-score
HPRD
ClusterONE	789	0.2307	0.1724	0.1973
MCODE	102	0.2059	0.0258	0.0458
MCL	1291	0.1255	0.1704	0.1445
CMC	44	0.3636	0.0178	0.0340
Coach	1762	0.2469	0.3890	0.3021
ProRank+	500	0.2820	0.1625	0.2062
PEWCC	1194	0.2739	0.2299	0.2499
SLPC only	2713	0.3693	0.4901	0.4212
d=32	858	0.7005	0.3785	0.4914
d=64	871	0.7107	0.3983	0.5105
d=128	841	0.7099	0.3890	0.5026
d=256	882	0.6961	0.3877	0.4980
d=512	823	0.7096	0.3831	0.4976
d=1024	867	0.7105	0.3970	0.5093
DIP
ClusterONE	363	0.5069	0.4012	0.4479
MCODE	82	0.0244	0.0030	0.0053
MCL	436	0.3463	0.3952	0.3692
CMC	262	0.4389	0.2912	0.3501
Coach	747	0.4351	0.5156	0.4719
ProRank+	167	0.4731	0.1516	0.2296
PEWCC	666	0.5916	0.3744	0.4586
SLPC only	1061	0.6447	0.4829	0.5522
d=32	719	0.8108	0.4428	0.5728
d=64	710	0.8070	0.4473	0.5755
d=128	702	0.8148	0.4368	0.5688
d=256	708	0.8263	0.4413	0.5753
d=512	711	0.8158	0.4413	0.5728
d=1024	691	0.8249	0.4354	0.5699

d denotes the dimension of each vector. No. of complexes denotes the total number of predicted complexes by each method. Bold value denotes the best score corresponding to F-score

ISSN: 1471-2105