Perturb Bio

Knock down a gene, or block a current. Watch the heartbeat change.

A cross-scale causal model of a human iPSC-cardiomyocyte. Knock down a gene with CRISPRi, or switch to drug-block mode and apply a percent block to an ion current. Either way the action potential is recomputed in real mV and ms by the published, validated Kernik-Clancy 2019 model. Each channel case moves the action potential in the direction the literature reports.

CRISPRi knockdown
baseline perturbed other knockdown levels
upstroke (zoom)
The first ~10 ms of the beat, foot-aligned, at the full 0.1 ms simulation resolution. The faint line is the baseline upstroke; the bright line is the current level. A slower upstroke (lower dV/dt max) is a shallower rising edge. This is a single iPSC-cardiomyocyte upstroke. It is not a tissue conduction reading or an ECG.
APD90 = action-potential duration to 90% repolarization. Cells beat spontaneously; no external pacing. Dashed lines mark peak, the 90%-repolarization level, and the maximum diastolic potential (MDP).
closed-loop discovery

The model also scores which experiment to run next. In a self-contained test it ranks all candidate knockdowns; a mechanistic prior finds every functionally active perturbation in far fewer experiments than random, while a DEG-count proxy (rank by how many genes move) does worse than random. This is expected: the loudest perturbations are ion channels, which are network leaves that move no other gene, so a DEG-count proxy ranks them last. Fewer experiments is better.

Ground truth here is the model's own forward simulation, so this shows the acquisition approach is internally consistent, not real-world predictive skill. That is a wet-lab question.
scope and limits

What it is: the cross-scale causal architecture running on a published, validated electrophysiology model (Kernik-Clancy 2019, via Myokit/CVODE). Its baseline reproduces the published biomarkers within ~1% (APD90 413.4 vs 414 ms), and each channel knockdown moves the action potential in the literature-consistent direction. A gene knockdown flows through a gene-regulatory network to ion-channel conductances to the action potential, in real units.

What it is not: the calibrated model. The gene-regulatory network here is a hand-seeded placeholder, and the gene-to-conductance step is an uncalibrated first-order map, not a per-gene fit. The functional layer is validated; the cross-scale composition is shown here and is testable against the program's CRISPRi data. The calibrated network and per-gene gains come from that data.

On drug-block mode: it scales one channel's conductance and shows how the single-cell action potential reshapes, the way an IKr or ICaL blocker would in this model. It is not a cardiotoxicity prediction and not a clinical ECG or QRS reading. This is a single iPSC-cardiomyocyte action potential, not tissue or a body-surface signal. It shows direction and rough magnitude of the AP change a current block produces in a validated model.

get in touch
daniel@perturb.bio