Table 1.

Prediction performances with different dataset sizes and angles on the DeepSnap-Deep Learning.

	Angles	120°		180°		240°		300°		360°
	tra:val:test	Means	SD	Means	SD	Means	SD	Means	SD	Means	SD
AUC	1: 1: 1	0.996	0.004	0.997	0.002	0.997	0.001	0.996	0.002	0.855	0.012
AUC	2: 2: 1	0.997	0.003	0.996	0.002	0.997	0.002	0.996	0.003	0.874	0.016
AUC	3: 3: 1	0.999	0.001	0.999	0.001	0.999	0.001	0.998	0.001	0.905	0.020
AUC	4: 4: 1	0.999	0.001	0.998	0.002	0.999	0.001	0.999	0.001	0.911	0.025
AUC	5: 5: 1	0.997	0.002	0.998	0.002	0.998	0.001	0.998	0.001	0.909	0.017
AUC	5: 5: 1 PMT	0.519	0.028	0.527	0.019	0.527	0.025	0.527	0.014	0.526	0.019
Acc (Test)	1: 1: 1	0.984	0.007	0.982	0.007	0.982	0.006	0.981	0.008	0.712	0.027
Acc (Test)	2: 2: 1	0.985	0.011	0.981	0.009	0.983	0.006	0.979	0.008	0.747	0.025
Acc (Test)	3: 3: 1	0.985	0.007	0.990	0.005	0.986	0.004	0.983	0.011	0.812	0.042
Acc (Test)	4: 4: 1	0.987	0.008	0.986	0.006	0.990	0.005	0.988	0.007	0.836	0.045
Acc (Test)	5: 5: 1	0.989	0.006	0.987	0.008	0.983	0.009	0.981	0.012	0.814	0.055
Acc (Test)	5: 5: 1 PMT	0.408	0.208	0.511	0.183	0.412	0.179	0.457	0.212	0.426	0.193
MCC	1: 1: 1	0.924	0.028	0.911	0.030	0.911	0.026	0.907	0.036	0.352	0.015
MCC	2: 2: 1	0.924	0.049	0.905	0.040	0.914	0.028	0.898	0.035	0.391	0.026
MCC	3: 3: 1	0.927	0.032	0.946	0.025	0.927	0.018	0.916	0.046	0.462	0.053
MCC	4: 4: 1	0.932	0.038	0.930	0.029	0.947	0.022	0.938	0.034	0.489	0.067
MCC	5: 5: 1	0.942	0.028	0.935	0.036	0.917	0.040	0.909	0.050	0.489	0.063
MCC	5: 5: 1 PMT	−0.038	0.044	−0.012	0.067	−0.015	0.068	−0.030	0.101	−0.009	0.079

Parameters (MPS:100, ZF:100, AT:23%, BR:14.5mÅ, BMD:0.4Å, BT:0.8Å, LR:0.0008, BS:108, GoogleNet).

5:5:1 PMT showed permutation test using Tra:Val:Test = 5:5:1.

ZF, zoom factor; AT, atom size; BR, bond radius; BMD, minimum bond distance; BT, bond tolerance; LR, learning rate; BS, batch size.