A fast mechanism answer, compared to the tools already on the market.

When a molecule shows a repolarization signal, a safety team answers one question first: which ion-current changes can produce it, and what would prove it. There are three standing ways to get that answer.

The free baseline is the CiPA in-silico TdP model. It is regulator accepted and costs nothing, but it runs on a healthy adult-ventricular cell and returns a risk class, not a ranked mechanism or a designed experiment.

The licensed simulator, most often Certara, is a platform you buy, learn and run. It is strong on PBPK and population spread. It is a tool, not an answer: you still must supply channel and exposure data and do the interpretation.

The wet CRO runs the measured assay on CiPA-qualified iPSC-cardiomyocytes. It gives real data at real cost, after you have decided what to test.

The read sits before all three. It takes the panel and exposure you already have, returns the ranked mechanism and the one experiment that would confirm it, and tells you which of the three doors, if any, you need.

CiPA is a regulatory success. The CiPAORdv1.0 model plus dynamic hERG binding classifies the healthy-cell TdP panel near perfectly, and we reproduce that: on the same 28-drug panel our iPSC-CM model reaches an AUC of about 0.91, below the published ORd/qNet figure of 1.00 and no better than a hERG-only margin on the high-risk call.

Where it wins: it is free, accepted, and better than us at classifying torsade risk in a healthy cell. On that task we do not compete with it, and we do not claim to.

Where it leaves a gap: it runs on one healthy adult-ventricular cell, so there is no disease-variant background to vary, and its output is a risk class. It does not rank the competing currents, and it does not return the experiment that would confirm the cause. That is the read.

Certara Cardiac Safety Simulator is the direct incumbent for in-silico proarrhythmia. It evaluates up to seven channel interactions, integrates PBPK, and predicts channel inhibition when in-vitro data are absent. Its population spread is parametric variability layered on a biophysical channel model, which is exactly the right tool for a dosing and population question.

Where it wins: PBPK, population variability, regulatory standing, and a validated parameter base. If your question is dose-to-concentration across a population, it is the tool, and a read is not a substitute for it.

Where it leaves a gap: it is a license you must learn and run, and it returns a number you must interpret. It does not name the one experiment that would confirm the mechanism, and no measured cell state enters the model. The read is the fast decision door before you commit to a platform: it tells you which currents to worry about and which experiment settles it, at a fraction of a license, in one to two weeks.

The wet labs run the real assay on CiPA-qualified iPSC-cardiomyocytes: MEA, calcium transients, impedance. They are the empirical ground truth, and they are the only place a mechanism claim becomes measured fact.

Where they win: measured data on a live cell. Nothing in-silico replaces that.

Where the read fits: before the wet run. The read ranks the competing currents and returns the specific assay that would discriminate them, so you test the informative thing first instead of screening blind. The confirmatory experiment the read names is often the SOW for the wet lab.

It is not a thorough QT study, it is not a CiPA submission, and it is not GLP. It does not replace the wet lab, and it is not a substitute for PBPK when the question is dosing.

A single-cell model is a proxy for tissue-level conduction, so the read bounds causes and ranks hypotheses. It does not prove causation. Kernik-2019 carries no distinct late-sodium current, so INaL block is invisible to it. The mechanism ranking is tested against phenotypes the same model generated; the direction of the underlying risk model is what the two measured validations check.

A team already holds these tools, and each is strong where it was built to work. The mechanism read is the cheaper first door: the decision of which experiment to run next, before the license and before the wet spend.