. 2017 Nov 14;8:816. doi: 10.3389/fphar.2017.00816

Table 4.

Binary TdP classifier scores of one-step and two-step classification on the direct features for the eight datasets.

Datasets			One step classifier	Two step classifier	Channel currents
Datasets	N_Drugs	Original model Accuracy (Feature)	C_drug = EFTPC	C_drug = IC_60,hERG
Dataset 1 (Mirams et al., 2011)	31	97 (APD₉₀)	90 (97)	100 (100)	I_Kr& I_CaV
Dataset 2 (Kramer et al., 2013)	55	91 ( $- l o g (\frac{I C_{50, h E R G}}{I C_{50, C a V}})$ )	89 (82)	95 (95)	I_Kr & I_CaV
Dataset 3 (Okada et al., 2015)	12	100 ( $\frac{C_{D r u g, A r r h y t h m i a}}{E F T P C}$ )	58 (66)	92 (92)	I_Kr & I_CaV
			66 (66^*)	100 (92^*)	I_Kr, I_CaV & I_NaL
Dataset 4 (Lancaster and Sobie, 2016)	86	87 (APD₅₀ & Diastolic Ca²⁺)	85 (85)	89 (90)	I_Kr & I_CaV
Dataset 5 (Crumb et al., 2016)	30		83 (83)	83 (83)	I_Kr & I_CaV
			80 (83^*)	80 (83^*)	I_Kr, I_Cav & I_NaL
Dataset 6 (Ando et al., 2017)	36	83 (APD prolongation and EAD (iPSCs))	86 (83)	86 (86)	I_Kr & I_CaV
Dataset 7 (Li et al., 2017)	12	100 $(\frac{A U C I_{N a L, d r u g}}{A U C I_{N a L, c o n t r o l}} + \frac{A U C I_{C a V, d r u g}}{A U C I_{C a V, c o n t r o l}})$	92 (92)	92 (92)	I_Kr & I_CaV
			100 (92^*)	100 (100^*)	I_Kr, I_CaV & I_NaL
Dataset 8	197		79 (79)	85 (86)	I_Kr & I_CaV

The total number of drugs in the particular dataset are reported in the N_Drugs column. Table also lists the feature used for classification in the original methods and corresponding scores if applicable. The last column shows a list of channel currents used to construct the classifiers. For comparision with the classification accuracy obtained with the direct features, we reported the highest accuracy obtained for the classifiers from the 13 derived features in isolation at different pacing rate/cell type in the bracket in One step and Two step classifier columns.

The derived features were extracted from simulations of the drug-induced block of all the reported ion channel currents in the in-vitro assay datasets.