A manually collected dataset that investigates the proarrhythmic risk of 109 drugs using parameters obtained from biophysical models. The dataset provides 4 torsade de pointes (TdP) risk categories. For all 109 drugs, IC50 values and Hill coefficients (h) for INa, INaL, IKr, Ito, ICaL, IK1, and IKs and human effective free therapeutic plasma concentration (EFTPC) were obtained (Llopis-Lorente et al., 2020). Task: The dataset can be used to study causal discovery methods aiming to infer ion channels selection for the drug-induced TdP risk. Summary: Size of dataset: 109 x 16 Task: Causal Discovery Problem Data Type: Mixed Data Dataset Scope: Standalone Dataset Ground Truth: Unknown Graph Temporal Structure: Static Data License: CC BY-NC 4.0 (see https://pubs.acs.org/page/policy/authorchoice_termsofuse.html) Missing Values: Existing Missing Values Missingness Statement: A dash in the dataset means that no value was found publicly available, thus it was considered that the drug effect on that current was negligible (Llopis-Lorente et al., 2020). Features: Class: Torsadogenic risk characterization. Class 1 (known risk of TdP), class 2 (possible risk of TdP), class 3 (conditional risk of TdP), and class 4 (drugs with a lack of evidence of TdP). EFTPC: Effective free therapeutic plasma concentration. EFTPC is the minimum concentration of the unbound form of a drug in plasma required to produce the intended therapeutic effect in the body. IC50[...]: Value that represents the minimal concentration of a drug that is required for 50% inhibition in vitrovalues (given in nm). Drugs are in the set {Kr, INa,INaL, ICaL, IKs, IK1, Ito}. h[...]: Hill coefficients of the currents. Drugs are in the set {Kr, INa,INaL, ICaL, IKs, IK1, Ito}. Files: med_file.csv: dataset