Table 4.

Evaluation results for four DO subnetworks

	Edge Correctness			Ancestor Correctness			F-score (Ancestor precision, ancestor recall)
	Parent			Parent			Parent
Root disease	Promotion	CliXO	MWST	Promotion	CliXO	MWST	Promotion	CliXO	MWST
Cardiovascular disease	0.06	0.09	0.07	0.32	0.18	0.11	0.50	0.27	0.21
							(0.57, 0.44)	(0.24, 0.30)	(0.13, 0.48)
Gastrointestinal disease	0.17	0.13	0.03	0.37	0.26	0.14	0.55	0.39	0.26
							(0.56, 0.53)	(0.36, 0.42)	(0.18, 0.48)
Musculoskeletal disease	0.16	0.08	0.10	0.15	0.26	0.09	0.26	0.41	0.17
							(0.44, 0.18)	(0.42, 0.40)	(0.16, 0.19)
Nervous System disease	0.13	0.07	0.09	0.29	0.17	0.10	0.46	0.30	0.19
							(0.70, 0.34)	(0.26, 0.34)	(0.13, 0.34)

Average Edge Correctness (EC), Ancestor Correctness (AC), Ancestor Precision (AP), Ancestor Recall (AR) and F-score across four DO subnetworks. Standard deviation is shown in parentheses. Best performance across different inference techniques is highlighted as italic