Table 1. Data sources of gene-disease associations.

Source	Version/Accessed	Type	#Associations	#Map MeSH	% Map MeSH
CTD [50]	Jul_02_2015/Aug’15	Curated	24,346	23,813	97.8
OMIM® [37]	18-08-2015/Aug’15	Curated	5,143⋆	3,375	65.6
Orphanet [40]	2015_07_31/Jul’15	Curated	6,094	1,744	28.6
UniProtKB [39]	2015_08	Curated	4,679	3,203	68.5
GWAS Catologue [43]◇	24_08_2015/Aug’15	Experimental	13,326	5,112	38.4
BeFree [8]	24-Aug-2015/Aug’15	Literature	330,888	233,264	70.5
GoF/LoF⊲	-/Oct’15	Literature	4,793	3,459	72.2
SemRep [41]∘	25/Feb’15	Literature	96,024	72,908	75.9
MGD [9, 11]*	24_08_2015/Aug’15	Predicted	1,943	1,577	81.2
RGD [10]*	21_08_2015/Aug’15	Predicted	7,667	7,667	100

Datasources used for G-D associations. ‘Curated’ refers to manually curated associations, ‘Experimental’ refers to associations drawn directly from genetic experimental observations, ‘Literature’ refers to associations automatically mined from literature and ‘Predicted’ refers to associations statistically inferred from animal models.

^⋆ Not including 1,397 associations for which the molecule basis is unknown.

^◇Threshold of 1e-7 was used.

^⊲See S1 Article.

^∘Extracted associations between gene and disease that were of the following predicates: AFFECTS; ASSOCIATED_WITH; AUGMENTS; CAUSES; PREDISPOSES; COEXISTS_WITH and NEG_ASSOCIATED_WITH as described in [8].

*Used the same parameters used by DisGeNET to extract predicted associations.