Fig 1.
Summary of the main keypoints of the article.
Fig 2.
Scheme of the materials and methods section.
Table 1.
Section Channel classification: Experimental data information.
IC50, Concentrations and Cmax are given in μM. T/V column corresponds to the compound embedding into the Training set (T) or Validation set (V).
Fig 3.
Section Heterogeneity: Finite element meshes of MEA used and example of heterogeneity field.
Left: Finite element mesh representing one well including 9 electrodes of the 6–well MEA device from Multichannel Systems (used in TdP classification). MEA device documentation is available on: http://www.qichi-instruments.com/bookpic/20163120452599.pdf. Right: Finite element mesh representing one well including 8 electrodes of the 96–well MEA device from Axion Biosystems with an example of generated cell heterogneity field (used in Channel classification). MEA device documentation is available on: https://www.axionbiosystems.com/sites/default/files/resources/mea_plates-brochure-rev_06.pdf.
Fig 4.
Section Drug modeling: Channel activity average and standard deviation for a Hill coefficient varying from 0.6 to 1.4.
The abscisse is the concentration factor with respect to the IC50.
Fig 5.
Section Drug modeling: EAD simulation.
Transmembrane action potential (AP, blue), extracellular field potential (FP, orange) and intracellular calcium transient trace (green) in a simulated EAD case.
Fig 6.
Section Dictionary entries: Electrophysiological biomarkers: List of the parameters computed on FP (up) and Calcium transient (down).
RC: Repolarization Center; FPD: Field Potentiel Duration; DA: Depolarization Amplitude; FPN: Field Potential Notch; AUCr: Area Under Curve of the repolarizaztion wave; RA: Repolarization Amplitude; RW: Repolarization Width; CA: Calcium Amplitude; DC: ‘Drowsing Calcium’; CDX: Calcium Duration. See Sections Field Potential Biomarkers computation and Calcium Signals Biomarkers computation (S1 Text).
Fig 7.
Section Dictionary entries: Electrophysiological biomarkers: Moxifloxacin simulation.
Simulation of the effect of Moxifloxacin at effective free therapeutic plasma concentration (10.96μM, see Table F in S1 Text) on the FP (from one electrode) and intracellular calcium transient (from one well) for two different heterogeneity fields. A finite element mesh of 96-well MEA device from Axion Biosystems was used for this simulation (see right panel of Fig 3).
Fig 8.
Section Dictionary entries: Electrophysiological biomarkers: Extended dictionary based on repolarization.
Upper panel: FP repolarization. Lower panel: Repolarization of cells affected by a compound with respect to the control case repolarization. The red line corresponds to the case where the repolarization is not affected.
Fig 9.
Dictionary entries: Wavelets: Signal reconstruction from wavelet coefficients.
Reconstruction of the absolute difference between the drug and control signals for the plateau and repolarization phases, based on wavelets coefficients.
Fig 10.
Section Classification optimization: Optimised algorithm in one dimension interpreted as a neural network.
Fig 11.
Section Classification optimization: Optimised algorithm for d dimensions interpreted as a neural network.
Fig 12.
Section TdP classification: TdP risk classification through simulations of 86 compounds.
Left: Validation versus Cost curve depending on the number of components and the dimension. Right: Drug repartition in the input space after convergence of the algorithm.
Fig 13.
Section TdP classification: Confusion matrices obtained for TdP risk classification of 86 compounds after convergence of the algorithm.
Yes: TdP risk. No: No TdP risk. Left: Training set (sample size: 1520) using randomized K-fold cross-validation. Sensitivity = 0.98, Specificity = 0.85 and Accuracy = 0.92. Right: Validation set (sample size: 200). Sensitivity = 1, Specificity = 0.675 and Accuracy = 0.935.
Fig 14.
Section Channel classification: Simulated FP under control and compound conditions.
FP trace from one electrode, showing the effect of drug simulation blocking the sodium channel at 4%, calcium channels at 3.6% and potassium channel at 27.9%.
Fig 15.
Section Channel classification (Binary classification part): Weights obtained by the optimised classification algorithm.
Fig 16.
Section Channel classification (Binary classification part): Experimental data classification in binary case.
Plain (resp. dotted) lines correspond to the average confidence of the LDA classifier for well classified (resp. misclassified) compound (well classification is according to Table 1). The black values on the lines correspond to the proportion of well classified observations for each compound.
Fig 17.
Section Channel classification (Binary classification part, Bepridil classification results): Example of experimental data with Bepridil, showing an increase in FPD and a decrease in DA of Pluricyte Cardiomyocytes.
Fig 18.
Section Channel classification, Binary classification part: Experimental data classification in binary case for each concentration.
Some concentrations were not used due to the quiescence or noisy signal observation. For each concentration, the LDA classifier returns the average probability for well classified (dotted bars) and misclassified (hatched bars) compounds.
Fig 19.
Section Ternary classification (Ternary classification part): Experimental data classification in ternary case.
Fig 20.
Section Ternary classification (Ternary classification part): Experimental data classification in ternary case for each concentration.
Fig 21.
Section Channel classification (Ternary classification part, Chlorpromazine classification results): Example of experimental FP trace with Chlorpromazine.