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

BMC Bioinformatics

Table 4 Experimental results obtained by using RF to filter the candidate complexes which are predicted from the modified DIP network by filtering edges with different simi-thres

Simi-thres	No. of edges left	No. of complexes	Precision	Recall	F-score	Δ
0.65	12167	653	0.8760	0.4413	0.5869	+0.0347
0.66	11941	667	0.8726	0.4428	0.5875	+0.0353
0.67	11683	652	0.8712	0.4368	0.5819	+0.0297
0.68	11423	634	0.8801	0.4324	0.5799	+0.0277
0.69	11174	617	0.8995	0.4294	0.5813	+0.0291
0.70	10946	612	0.9020	0.4235	0.5764	+0.0242
0.71	10673	610	0.8918	0.4235	0.5743	+0.0221
0.72	10410	616	0.8929	0.4235	0.5745	+0.0223
0.73	10184	622	0.8939	0.4264	0.5774	+0.0252
0.74	9907	608	0.8947	0.4160	0.5680	+0.0158
0.75	9633	594	0.9091	0.4190	0.5736	+0.0214

Δ denotes the improvement of F-score compare with using SLPC alone. Bold values denote the best scores corresponding to the specific metric

ISSN: 1471-2105