Structure-Informed In Silico Stress Testing of Latent Ion-Current Triggers in Developmentally Gated SADS Vulnerability

Short title: Latent Current Trigger Stress Testing

This repository defines a conservative, hypothesis-generating computational framework for testing whether a structurally plausible but weakly expressed or conditionally gated hypothetical current (I_new) could be electrophysiologically silent at baseline yet become destabilizing under mutation-specific gating shifts, calcium stress, reduced repolarization reserve, pacing perturbation, or a computational developmental transition.

The project is not a clinical decision tool and does not claim to predict sudden death, prove a human embryonic mechanism, establish that a candidate protein is definitely an ion channel, or infer safety from APD normalization.

One-Page Project Summary

Rare SADS-associated perturbation pairs and early human developmental cardiac electrophysiological states occupy an ethically inaccessible state space: they cannot be systematically provoked in humans, and early embryonic-like human cardiac states cannot be directly stress-tested as if they were adult cardiomyocytes. Immature cardiomyocytes also differ from adult cardiomyocytes in ion-channel expression, resting membrane potential, automaticity, calcium handling, and stress responses, so a single adult action-potential model is an incomplete proxy.

This project creates a structure-informed family of plausible I_new models rather than a single asserted channel mechanism. Candidate protein features such as transmembrane domains, pore-like motifs, voltage-sensor-like charged residues, calcium-binding motifs, regulatory domains, disease-associated variants, cardiac/developmental expression, residue conservation, or possible accessory-channel behavior are used only to motivate parameter priors. They do not uniquely determine conductance, kinetics, reversal potential, or gating.

The framework embeds I_new into a human ventricular action-potential model, with ORd/O'Hara-Rudy as the intended full baseline and a simplified minimal proxy for the runnable prototype. It then samples virtual cells across baseline-current variability, I_new parameter uncertainty, mutation transformations, stress protocols, and a maturation axis m = 0..1 from immature-like to adult-like states. It quantifies stress-revealed instability using EAD, DAD, spontaneous action potential, failure-to-repolarize, alternans, restitution, and calcium-transient metrics.

The key computational question is: can a latent, structure-informed current family generate a reproducible region of parameter space in which baseline biomarkers remain apparently compensated while stress, mutation, or maturation unmasks triggered instability?

Claim Boundaries

Established biological facts used as constraints:

• Human ventricular action-potential models such as ORd are established tools for mechanistic cellular electrophysiology.
• hiPSC-CMs and immature cardiomyocyte-like systems differ from adult cardiomyocytes in electrophysiology and calcium handling.
• EADs occur during repolarization, DADs occur after repolarization, and reduced repolarization reserve can promote afterdepolarizations.
• Population-of-models and CiPA-style in silico workflows are accepted research paradigms for variability and drug-safety modeling.

Plausible modeling assumptions:

• Structural/domain evidence can justify a family of hypothetical current models.
• Developmental maturation can be represented by a proxy axis that changes conductances, calcium handling, capacitance, automaticity, and stress response.
• Mutations are parameter transformations sampled from priors, not fixed deterministic truths.

Speculative hypotheses:

• Some I_new variants may be silent at baseline but stress-revealed.
• Some mutant I_new regimes may be tolerated in immature-like electrophysiology but destabilize during maturation.
• I_new may modify known SADS-associated backgrounds by destroying apparent compensation under stress.

Computational outputs that would support the hypothesis:

• High concealed-trigger index: normal-looking baseline biomarkers plus stress-induced EAD/DAD/spontaneous AP.
• Increased trigger activation probability for mutant I_new versus null/WT controls.
• A developmental unmasking pattern: stability at low m, instability at higher m.

Outputs that would weaken the hypothesis:

• Mutant I_new produces instability only by grossly abnormal baseline APs.
• Null, WT, and mutant groups have indistinguishable stress responses across broad ensembles.
• Results depend on a tiny, non-robust parameter corner or numerical artifacts.

Quick Start

cd C:\Users\User\latent-current-trigger
python src\run_simulation.py --n 100 --seed 7

The toy prototype writes only under results\toy_prototype\.

Real-Data-Constrained Workflow

cd C:\Users\User\latent-current-trigger
python src\run_real_data_constrained.py --n 1000 --seed 20260505

This workflow keeps outputs separated:

• results\toy_prototype\: legacy/minimal toy prototype outputs.
• results\real_data_constrained\: synthetic simulations constrained by downloaded CellML models and curated public data.
• results\model_validation\: baseline traces/metrics from downloaded CellML-generated model implementations.
• results\provenance\: source audit, model sources, and candidate-protein evidence audit.

Minimal Prototype Scope

The current runnable prototype is intentionally modest. It uses a Mitchell-Schaeffer-style minimal action-potential proxy with a calcium state variable and a structure-informed voltage/Ca-sensitive I_new. It is suitable only for pipeline validation: sampling, perturbing, pacing, stress testing, detecting events, computing scores, and generating figures. It is not an ORd replacement.

Must-Cite Literature

• O'Hara et al. 2011, ORd human ventricular AP model: https://doi.org/10.1371/journal.pcbi.1002061
• Britton et al. 2013, experimentally calibrated populations of models: https://doi.org/10.1073/pnas.1304382110
• Passini et al. 2017, human in silico drug trials: https://doi.org/10.3389/fphys.2017.00668
• Li et al. 2018, CiPA in silico proarrhythmia model: https://doi.org/10.1002/cpt.1184
• Paci et al. 2013, hiPSC-CM computational models: https://doi.org/10.1007/s10439-013-0833-3
• Paci et al. 2012, hESC-CM AP modeling: https://doi.org/10.1186/1475-925X-11-61
• Keung et al. 2014, hPSC-CM maturation review: https://doi.org/10.1186/scrt406
• Weiss et al. 2010, EAD mechanisms review: https://pmc.ncbi.nlm.nih.gov/articles/PMC3005298/
• Catterall 2010, voltage-sensor structure/function: https://doi.org/10.1016/j.neuron.2010.08.021
• ICH E14/S7B Q&A and FDA guidance: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/e14-and-s7b-clinical-and-nonclinical-evaluation-qtqtc-interval-prolongation-and-proarrhythmic
• ISSCR stem-cell research guidelines: https://www.isscr.org/stem-cell-guidelines

Repository Layout

latent-current-trigger/
  configs/        model, mutation, developmental, stress, and ensemble specs
  src/            runnable minimal prototype plus planned ORd/hiPSC integration modules
  results/        raw traces, processed tables, figures
  paper/          project outline, abstract, methods scaffold
  notebooks/      planned notebook workflow