This post is to announce that we are making our in-silico action potential prediction web portal (AP-predict) available for public use.
The portal is free, you just need to register for an account. We’d be happy to have any feedback on it, to make improvements.
What does it do?
The portal allows you to integrate information on multiple ion channel screens, using a range of mathematical action potential models (taken from the CellML repository), to see a compound’s expected effect on the whole cellular action potential, as suggested as part of the CiPA initiative. The portal is intended to make these electrophysiology simulations accessible to safety pharmacology teams.
The portal takes as its input data 50% Inhibitory Concentrations (IC50 – the concentration of a compound which blocks [peak] ion channel current by 50%); for any/each of the commonly-affected cardiac ion channels. The IC50 values are used to calculate a percentage ion channel block at a particular concentration, for each ion channel of interest (by assuming a standard Hill curve describes the full concentration-effect relationship). These percentage blocks are then used to reduce the maximum conductances of the ion channels, and the action potential model is paced to steady state to simulate the effect of the compound on a whole cardiac cell. The process is repeated for a range of concentrations; building up an action potential duration vs concentration curve.
How does it work?
The portal provides a simple interface to the ApPredict simulation code (a Chaste bolt-on project) which is a C++ command-line program, running on a server here in Oxford. If you are interested you can install it yourself which provides a bit more flexibility, but this web portal provides a simple interface to the most widely used components. The portal also stores the simulation inputs and outputs for easy retrieval later on, and plots the outputs for you. You can download or view the source code for the portal too, which could be useful if you want to run it on your own system e.g. inside a firewall for data security.
Evaluation and intended use
We have written a user guide which can be found on the About page of the portal, which is recommended reading. We have published a series of papers evaluating the predictive power of these simulations, based on various ion channel IC50 datasets. For more information see the user guide and my list of publications.