Table 1.

A subset of negative-class proteins predicted to be disease-related

Conf. Score	OS	DORIF	OMIM	Suspected Disease Relationship
6.24096	CDH5	-	-	Melanoma, tumor metastasis
5.8186	PTCH1	-	109400, 605462, 610828	Basal Cell Nevus Syndrome, Basal Cell Carcinoma
5.19721	STAMBPL1	-	-	Very light evidence, Alzheimer's
5.14813	MDH2	-	-	Very light evidence, tumor development
1.09972	DPP4	-	-	Diabetes (17 PMIDs), colon cancer (3 PMIDs)
1.09972	GRK5	-	-	Very light evidence, heart failure
0.907016	GZMB	-	-	Lymphoma (30 PMIDs), tumors (92 PMIDs)
0.898631	TCF4	-	610954	Pitt-Hopkins Syndrome, various cancer (light evidence)
0.705929	FGR	-	-	Breast cancer (3 PMIDs), prostate cancer (1 PMID)
0.705929	FLT1	-	-	Cancer, various
0.705929	PECAM1	-	-	Cancer, various
0.705929	SREBF2	-	-	Prostate cancer (2 PMIDs)
0.705929	STAT6	-	-	Prostate cancer (3 PMIDs)
0.705929	TOP1	-	-	Leukemia, colon and ovarian cancer
0.664823	CD74	-	-	Very light evidence, lymphoma

A subset of 15 proteins belonging to the non-disease-related class (lacking DORIF annotation) but predicted to be disease-related, sorted by the ADTree-assigned confidence score. Two proteins (PTCH1 and TCF4) have associated OMIM disorders. 9/15 proteins (CDH5, DPP4, GZMB, FGR, FLT1, PECAM1, SREBF2, STAT6, and TOP1) have moderate to strong evidence of disease association, while 4/15 (STAMBPL1, MDH2, GRK5, and CD74) have light evidence linking them to disease. (n PMIDs) indicates the number of PubMed IDs connecting a protein to a particular disease. 'Conf. Score' is the confidence score assigned by the ADTree classifier, 'OS' is the official symbol of the gene, 'DORIF' is Disease Ontology + Gene Reference Into Function, 'OMIM' is the MIM number associated with the gene, 'light evidence' is defined as having a predicted disease association according to the MalaCards database [48]. Disease information for this table was acquired from the GeneCards database [49].