OptNCMiner: a deep learning approach for the discovery of natural compounds modulating disease-specific multi-targets

Table 1 Data used to construct the base dataset, transfer learning dataset, and few-shot learning dataset

Dataset	Target gene	Target protein	Data source	Active compounds	Inactive compounds
Base dataset (actives \(>\) 5000)	ADORA2A	Adenosine receptor A2a	ExCAPE-DB	5077	591
	BRCA1	Breast cancer type 1 susceptibility protein	ExCAPE-DB	8619	43,095
	CNR1	Cannabinoid receptor 1	ExCAPE-DB	5125	397
	DRD2	D(2) dopamine receptor	ExCAPE-DB	8037	40,185
	HTR1A	5-hydroxytryptamine receptor 1A	ExCAPE-DB	6339	31,695
	KCNH2	Potassium voltage-gated channel subfamily H member 2	ExCAPE-DB	5327	26,635
	LMNA	Prelamin-A/C	ExCAPE-DB	14,533	72,665
	OPRM1	Mu-type opioid receptor	ExCAPE-DB	5665	2872
	SLC6A4	Sodium-dependent serotonin transporter	ExCAPE-DB	6912	370
	TARDBP	TAR DNA-binding protein 43	ExCAPE-DB	12,193	60,965
	TDP1	Tyrosyl-DNA phosphodiesterase 1	ExCAPE-DB	23,129	115,645
Transfer learning dataset (1000 > actives > 500)	ADRA2A	Alpha-2A adrenergic receptor	ExCAPE-DB	816	39
	GRIN1	Glutamate receptor ionotropic	ExCAPE-DB	553	92
	HTR3A	5-hydroxytryptamine receptor 3A	ExCAPE-DB	565	65
	MINK1	Misshapen-like kinase 1	ExCAPE-DB	929	8
	PKM2	Pyruvate kinase PKM	ExCAPE-DB	546	2730
	POLK	DNA polymerase kappa	LIT-PCBA	772	3860
	VDR	Vitamin D3 receptor	LIT-PCBA	884	4420
Few-shot learning dataset (100 > actives)	ADRB2	Beta 2 adrenergic receptor	LIT-PCBA	17	170
	ESR	Estrogen receptor alpha	LIT-PCBA	13	130
	IDH1	Isocitrate dehydrogenase	LIT-PCBA	39	390
	MTOR	mammalian target of rapamycin complex 1	LIT-PCBA	97	970
	OPRK1	Kappa opioid receptor	LIT-PCBA	24	5460
	PPARG	Peroxisome proliferator-activated receptor gamma	LIT-PCBA	27	270
	TP53	Cellular tumor antigen p53	LIT-PCBA	79	790

ISSN: 1471-2105