Machine learning to predict metabolic drug interactions related to cytochrome P450 isozymes

Table 7 The scaffold analysis results of the substrate and inhibitor datasets

Subsets			Number of scaffolds	Number of carbons	Most prominent scaffolds (Murcko Class)	Common scaffolds
1A2	Inhibitor	Positive	3520	1418	1526, 3567, 3606	4, 1, 168
	Inhibitor	Negative	4257	2090	54, 7220, 5, 414, 849	4, 1, 168
	Substrate	Positive	150	94	36, 8, 14	3, 56
	Substrate	Negative	163	113	158, 43	3, 56
2C9	Inhibitor	Positive	3088	1708	4866, 4863, 4885	4, 3366, 179
	Inhibitor	Negative	4353	1793	59, 3303, 3413	4, 3366, 179
	Substrate	Positive	168	119	66, 36, 70	19
	Substrate	Negative	166	123	173, 199, 254	19
2C19	Inhibitor	Positive	4189	2015	3756, 188, 349, 5176	4, 6, 39
	Inhibitor	Negative	3751	1614	3567, 461, 3570, 7237	4, 6, 39
	Substrate	Positive	132	89	7, 31, 63	47, 16
	Substrate	Negative	137	105	93	47, 16
2D6	Inhibitor	Positive	1836	1003	74, 3742, 3904, 3911	4, 36, 1655
	Inhibitor	Negative	6244	2580	3718, 3744, 56	4, 36, 1655
	Substrate	Positive	209	150	40, 42, 45	1, 26
	Substrate	Negative	223	153	212, 62, 38	1, 26
3A4	Inhibitor	Positive	3536	1912	3504, 3539, 3592	4, 3576, 168
	Inhibitor	Negative	4227	1630	58, 429, 7136	4, 3576, 168
	Substrate	Positive	518	385	57, 31, 80	38, 8, 60
	Substrate	Negative	506	301	541, 530, 533, 105, 631	38, 8, 60

ISSN: 1758-2946