Before you commit wet-lab time, run it through a validated human iPSC-cardiomyocyte model in silico. You get a TdP-risk call from your ion-channel panel, the currents behind it, the dose-response, and the experiment that would settle it. Days, no assay to stand up.
what we offer
in-silico cardiac read
A fast read on your compound's cardiac signal (QT, QRS, or repolarization change) on a validated iPSC-cardiomyocyte action-potential model (Kernik-Clancy 2019) with a virtual cell population. You get the risk call, the dose-response and reversibility, and the assay that would confirm it. No wet lab. Days to a first read.
ion-channel panel read
Bring your ion-channel panel (hERG, Nav1.5, Cav1.2). We run it through the same iPSC-cardiomyocyte model, as the CiPA workflow does, and return a TdP-risk class alongside the currents driving it. Validated on the 28-drug CiPA panel; see the worked examples.
co-designed wet confirmation
When the read points to a wet answer, we design the iPSC-cardiomyocyte pilot around your question and run it through partner labs (paired morphology and transcriptomics, or functional electrophysiology). You keep your results.
Cardiac safety is moving to human iPSC-cardiomyocytes and in-silico models, the path the FDA's 2025 new-approach-methods roadmap prioritizes. We start in silico, on a validated cardiomyocyte model, so you get a mechanism read in days without standing up an assay. When a read needs empirical confirmation, we design the wet pilot around your question and run it through partner labs. Public iPSC-CM cardiotoxicity data is narrow: the widely-cited sets cover a single drug (Burridge 2016, doxorubicin) or a single drug class (Sharma 2017, tyrosine-kinase inhibitors), and none pairs morphology with transcriptomics across a broad panel. We start in silico, and generate the confirmatory wet data only when a read calls for it.
what a read delivers
what's in a read
The risk call and the currents behind it: the TdP-risk class your panel maps to, and the ion-current changes (for example hERG/IKr, Nav1.5, or Cav1.2) driving it. Where more than one current story fits the signal, we say so, and we rank them against each other.
Dose and reversibility: the dose-response and washout behavior the model predicts.
The confirming experiment: the single assay that would settle the question, from an identifiability analysis of the model.
Limits of the read: a single-cell model is a proxy for tissue-level conduction, so the read bounds the causes rather than proving one, and states per current what it can and cannot resolve.
Your compound's signal in, a mechanism read out on the validated iPSC-CM model. Days, no wet lab.
->
if needed
Confirmatory design
The identifiability analysis names the one assay that would confirm the cause.
->
weeks
Co-funded wet pilot
We design and broker the iPSC-CM pilot through partner labs. You keep your results.
inside the wet pilot, when it runs
01 design
Scoped design
The pilot is designed around the confirmatory question from the in-silico read.
02 execute
Partner lab
The pilot runs at a partner CRO. Perturb designs it and reads the result.
03 measure
Multimodal
Cell Painting and DRUG-seq, co-measured on matched parallel arms (morphology + transcriptomics).
04 analyze
Mechanism
Results are read against the model to confirm or refine the mechanism.
-> The read designs the pilot; the pilot confirms the read.
The read designs the pilot. The in-silico read names the confirmatory question, so the wet pilot targets the mechanism rather than running a broad screen.
Daniel Reda. Two prior exits in life science data: CureTogether (acquired by 23andMe) and Redasoft (acquired by Hitachi). Background in Molecular Genetics.