BMC Bioinformatics

Table 5 The AUCs values of different graph convolutional networks methods over 16 diseases for threshold = 10

From: A novel candidate disease gene prioritization method using deep graph convolutional networks and semi-supervised learning

Disease	Methods
Disease	GCNGP	GCAS	RGCN	PGCN	GCN-MF
Pancreatitis	93.65	90.44	88.18	89.05	89.25
Parkinson’s disease	88.82	87.29	84.75	83.69	85.90
Celiac disease	87.39	83.08	82.77	80.88	85.10
Atherosclerosis	96.55	92.97	89.35	91.64	94.89
Esophageal cancer	87.44	85.15	84.58	81.29	86.96
Crohn’s disease	82.73	81.36	78.84	80.01	80.52
Breast cancer	87.66	84.49	81.69	83.13	85.73
Alzheimer’s disease	90.45	85.25	83.86	82.49	88.27
Ulcerative colitis	79.36	75.77	72.61	74.16	77.07
Endometriosis	96.05	95.29	91.54	89.99	92.43
Cirrhosis	74.48	71.01	69.37	70.71	72.55
Myocardial infarction	91.69	86.82	82.95	85.36	89.18
Tuberculosis	95.35	94.66	91.62	92.13	93.69
Lymphoma	92.81	88.91	86.25	85.79	90.14
Rheumatoid arthritis	84.97	83.54	80.08	81.22	82.25
Asthma	86.19	83.09	82.92	83.88	84.24
Average	88.47	85.57	83.21	83.46	86.14

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com