TABLE 3.

Stepwise results of the pre-processing procedure.

	Raw	Step 1	Cleaned	Step 2	Nodes
Disease	31,974	Removed noisy concepts like “cardioembolic”, “JAGS”, “nonvitamin”, etc. that could not be mapped to MeSH	31,963	MeSH	801
Chemical	4,494		3,724		678
Gene	11,211	Excluded genes that do not belong to Homo-sapiens	8,781	NCBI Gene	968
Gene variant
DNA mutation	69	Removed variants with unclear loci (i.e., could not be mapped to an SNP ID)	17	dbSNP	126
Protein mutation	349		91
SNP	104	—	104
Total	48,201	—	44,680	—	2,573