Table 4. Summary of All of the Data Sets Used in the Study^a.

		proteins	compounds	interactions
training	KIBA (IC50)b	961	30 474	61 624
validationc (DTI-MLCD)	enzyme	1411	1777	7371
	GPCR	156	1680	5383
	nuclear receptors	33	541	886
	ion channel	204	210	1476

^aKIBA data set was employed for training and internal validation, while the gold-standard data set from DTI-MLCD was used for external validation.

^bProteins for drugs listed in the KIBA data set were extracted manually from ChEMBL.

^cUsed as an external validation data set.