DCIMaL (Drug Classification and Identification by Modeling and Learning) is a Python and C++ software for safety pharmacology studies and particularly field potentials signals measured with micro-electrode array (MEA). This software is arranged in three parts:
- Field potential simulations:
It consists of a solver of Bidomain equations (which models the electrical activity of the heart) coupled with an action potential model (electrical activity of one cardiac cell) applied to a finite element mesh of a MEA. The resulting extracellular potential is coupled with an imperfect electrode model which leads to the field potential on the electrode. The impact of a chemical compound over the field potential can be simulated tuning one or some parameters of the action potential model.
- Dictionary extraction:
Once simulated (or experimental) field potentials are available, some quantities may be automatically extracted. In this part of the software, amplitudes, durations, wavelets coefficients,… are extracted in an automated way.
- Classification:
This part of the software is dedicated to the compound classification based on a linear combinations of the dictionary entries as features.
Developers:
- J-F. Gerbeau
- D. Lombardi
- F. Raphel
Gallery:
Field potential simulations:
Example of a MEA device with 9 electrodes. a) Finite element mesh of the MEA. b) Physical scheme of the same MEA.
Simulated potentials at electrodes. Using Paci action potential model.
Simulated Early After Depolarization. Using O’Hara Rudy action potential model.
Dictionary extraction:
Example of extracted dictionary entries.
Classification:
Compound classification. Channel blockade classification of 7 compounds based on experimental data.
