Computational models of cardiac disease — landscape survey
Provenance/reference document for dismech curation of computational_models:
blocks on cardiac disorder entries. Compiled July 2026. All PMIDs below were
verified against PubMed records or the publisher page; the four flagged as
"search-confirmed" were taken from search-result summaries with matching PubMed
links but not individually re-opened — double-check before hard-committing them
to a KB entry.
Why this matters for dismech
Cardiac electrophysiology is the most model-rich subfield of physiology, and the
dismech schema already supports it: the top-level computational_models: slot
(ComputationalModel class) captures ODE/PDE and other in-silico models, with
model_type: KINETIC for ODE systems, variables (each ModelVariable with a
native dataset_identifier, unit, and ontology mappings_list), and
ModelVariableDescriptor thresholds/severity_scale that link a state-variable
value to an HP phenotype. Evidence uses evidence_source: COMPUTATIONAL. The
prior worked example was CKD-Mineral_Bone_Disorder.yaml (Peterson-Riggs ODE
model, BIOMD0000000613); Long_QT_Syndrome.yaml is now the first cardiac
worked example (ORd, CiPA IKr-dynamic ORd, ToR-ORd).
Models organized by spatial scale
Cardiac modeling is canonically multiscale: ODEs describe subcellular fluxes and single-cell membrane dynamics; those cell models become the reaction term inside tissue/organ reaction-diffusion PDEs; continuum finite-element mechanics (plus fluid-structure interaction for blood) closes the electromechanical loop.
Subcellular / organelle (ODE systems)
| Sub-system | What is modeled | Representative models |
|---|---|---|
| Mitochondrial bioenergetics | TCA cycle, oxidative phosphorylation, ATP/ADP, NADH, membrane potential, matrix Ca²⁺ | Cortassa–Aon–Marbán–Winslow–O'Rourke 2003; ECME extension (Cortassa 2006) |
| Ca²⁺ handling / SR | L-type Ca²⁺ current, RyR2 Ca²⁺-induced Ca²⁺ release, SERCA uptake, compartmental Ca²⁺ | Shannon–Bers 2004 (reference SR/Ca²⁺ framework) |
| Myofilament / contraction | Cross-bridge cycling, Ca²⁺–troponin binding, length/tension dependence | Rice et al. 2008; Land et al. (used with ToR-ORd) |
These ODE sub-models are usually plugged into a single-cell electrophysiology model rather than run alone.
Single-cell cardiomyocyte electrophysiology (Hodgkin–Huxley lineage)
Systems of ODEs for membrane potential + gating variables + ion concentrations:
- Hodgkin–Huxley 1952 (squid axon) — mathematical template, not cardiac.
- Noble 1962 — first cardiac model (Purkinje fibre).
- Beeler–Reuter 1977 — first mammalian ventricular AP.
- Luo–Rudy I (1991) / II dynamic (1994) — guinea-pig ventricle; 1990s workhorse.
- Courtemanche 1998 / Nygren 1998 — foundational human atrial cell models (AF substrate).
- Shannon–Bers 2004 — rabbit ventricle, gold-standard Ca²⁺ handling.
- ten Tusscher 2004 / ten Tusscher–Panfilov 2006 — first widely adopted human ventricular models, efficient enough for tissue/organ use.
- Grandi–Pasqualini–Bers 2010 — human ventricular AP + Ca²⁺ transient; strong for CaMKII / heart-failure remodeling.
- O'Hara–Rudy "ORd" 2011 — modern reference human ventricular model; CiPA consensus model.
- ToR-ORd (Tomek et al. 2019) — ORd refinement (corrected reversal potentials, reformulated I_CaL, replaced hERG), validated in disease and drug block, cell-to-ECG.
- Paci et al. hiPSC-CM models (2013 →) — human iPSC-derived cardiomyocyte models (the main in-vitro human platform).
Key ventricular currents: I_Na, I_CaL, I_Kr (hERG — central to drug safety), I_Ks, I_K1, I_to, I_NaCa, I_NaK, plus SR fluxes and CaMKII regulation.
Tissue / organ (reaction-diffusion PDEs)
- Monodomain — single reaction-diffusion PDE (cell ionic model = reaction term + diffusion term); cheaper, assumes equal anisotropy ratios.
- Bidomain — coupled intra/extracellular PDE pair; required for defibrillation, external stimulation, ECG/body-surface potentials, unequal anisotropy.
- Whole-heart / organ — 3D geometries + fiber orientation for arrhythmia mechanism, ablation planning, in-silico drug trials (reviewed in Clayton 2011).
Mechanics / finite-element and fluid-structure
- Nonlinear anisotropic hyperelastic constitutive laws (Guccione, Holzapfel–Ogden) + active tension → electromechanical whole-heart models.
- Fluid-structure interaction couples deforming myocardium/valves to intracavitary blood flow (Navier–Stokes). Basis of "cardiac digital twin" pipelines.
Mapping model types to diseases (and candidate dismech entries)
| Disease | Typical modeling approach | dismech entry |
|---|---|---|
| Long QT syndrome | Ventricular ODE (ORd/ToR-ORd/Grandi) with I_Kr↓ (LQT2), I_Ks↓ (LQT1), late I_Na↑ (LQT3) → APD↑ → EADs | Long_QT_Syndrome ✅ (done) |
| Drug-induced TdP | CiPA IKr-dynamic ORd + multi-channel pharmacology → qNet | Long_QT_Syndrome ✅, Torsade_De_Pointes_Syndrome_With_Short_Coupling_Interval |
| Brugada syndrome | Ventricular cell + tissue, reduced I_Na (SCN5A) / altered I_to, transmural heterogeneity | Brugada_Syndrome |
| Short QT syndrome | Ventricular ODE with gain-of-function K⁺ currents → APD↓ | Short_QT_Syndrome |
| CPVT | Ca²⁺-handling models (Shannon–Bers/Grandi) with RyR2/CASQ2 leak → DADs | RYR2_CPVT |
| Atrial fibrillation | Courtemanche/Nygren atrial cell → 2D/3D tissue; rotor/ablation studies | Atrial_Fibrillation, Familial_Atrial_Fibrillation |
| HCM / DCM | Electromechanical FEM + cell remodeling (ToR-ORd validated in HCM) | Hypertrophic_Cardiomyopathy, Dilated_Cardiomyopathy |
| Heart failure | Cell remodeling (Grandi/Shannon–Bers) + organ electromechanics | Heart_Failure |
| MI / ischemia | Tissue reaction-diffusion with regional hyperkalemia/acidosis, scar geometry | Myocardial_Infarction |
The CiPA initiative
CiPA (Comprehensive in vitro Proarrhythmia Assay; FDA/HESI/CSRC) replaces the hERG-plus-QT paradigm with mechanistic in-vitro ion-channel data + an in-silico cell model + hiPSC-CM confirmation. The consensus in-silico model is the ORd model with a Markov IKr model capturing dynamic drug–hERG binding ("IKr-dynamic ORd"), optimized in Dutta 2017; the qNet metric stratifies high/intermediate/low TdP risk.
Repositories, databases, and software
| Resource | What it is | Identifiers | Landmark |
|---|---|---|---|
| CellML / Physiome Model Repository (PMR2) | Curated CellML cell/tissue models; hosts nearly all named cardiac cell models | models.cellml.org/e/<id> exposures (ORd = e/71, ToR-ORd = e/5f1) |
Yu 2011 (PMID:21216774) |
| BioModels | EBI repository (mostly SBML) | BIOMD0000000xxx (curated) |
Malik-Sheriff 2020 (PMID:31701150) |
| openCARP | Open-source cardiac EP simulator (mono/bidomain); successor to CARP | opencarp.org | Plank 2021 (PMID:34171774) |
| Chaste | C++ library for cardiac EP; strong CellML integration | github.com/Chaste | Mirams 2013 (PMID:23516352) |
| FDA/CiPA | Reference implementation of the IKr-dynamic ORd + qNet pipeline | github.com/FDA/CiPA | Dutta 2017 (PMID:28878692) |
| ToR-ORd | Model code (MATLAB + CellML), validation pipeline | github.com/jtmff/torord | Tomek 2019 (PMID:31868580) |
| Cardiac Atlas Project | Imaging + statistical shape/motion atlases of normal/pathological hearts | cardiacatlas.org | Fonseca 2011 (PMID:21737439) |
Identifier conventions: BioModels BIOMD0000000###; Physiome/CellML
human-readable exposure URIs (models.cellml.org/e/<id>). dismech's schema
declares a biomodels: prefix; there is no physiome:/cellml: prefix yet, so
CellML models are referenced by full repository_url.
Landmark papers with verified PMIDs
Single-cell electrophysiology lineage
| Model | Citation | PMID |
|---|---|---|
| Noble Purkinje | Noble D. J Physiol. 1962;160(2):317–352. | 14480151 |
| Beeler–Reuter | Beeler GW, Reuter H. J Physiol. 1977;268(1):177–210. | 874889 |
| Luo–Rudy I | Luo CH, Rudy Y. Circ Res. 1991;68(6):1501–1526. | 1709839 |
| Luo–Rudy II (dynamic) | Luo CH, Rudy Y. Circ Res. 1994;74(6):1071–1096. | 7514509 |
| Human atrial (Courtemanche) | Courtemanche M, Ramirez RJ, Nattel S. Am J Physiol. 1998;275(1):H301–H321. | 9688927 (search-confirmed) |
| Human atrial (Nygren) | Nygren A, et al. Circ Res. 1998;82(1):63–81. | 9440706 (search-confirmed) |
| Shannon–Bers Ca²⁺ | Shannon TR, et al. Biophys J. 2004;87(5):3351–3371. | 15347581 |
| ten Tusscher 2004 | ten Tusscher KHWJ, et al. Am J Physiol Heart Circ Physiol. 2004;286(4):H1573–H1589. | 14656705 |
| ten Tusscher–Panfilov 2006 | Am J Physiol Heart Circ Physiol. 2006;291(3):H1088–H1100. | 16565318 |
| ten Tusscher 2006 (efficient) | Phys Med Biol. 2006;51(23):6141–6156. | 17110776 |
| Grandi–Pasqualini–Bers | J Mol Cell Cardiol. 2010;48(1):112–121. | 19835882 |
| O'Hara–Rudy (ORd) | O'Hara T, Virág L, Varró A, Rudy Y. PLoS Comput Biol. 2011;7(5):e1002061. | 21637795 |
| ToR-ORd | Tomek J, et al. eLife. 2019;8:e48890. | 31868580 |
hiPSC-CM models (Paci)
| Citation | PMID |
|---|---|
| Paci M, et al. Ann Biomed Eng. 2013;41(11):2334–2348. | 23722932 |
| Paci M, et al. Biophys J. 2020;118(10):2596–2611. (all-optical) | 32298635 (search-confirmed) |
CiPA / drug-induced arrhythmia
| Citation | PMID |
|---|---|
| Sager PT, et al. Am Heart J. 2014;167(3):292–300. (CiPA rationale) | 24576511 |
| Colatsky T, et al. J Pharmacol Toxicol Methods. 2016;81:15–20. (CiPA update) | 27282641 |
| Dutta S, et al. Front Physiol. 2017;8:616. (IKr-dynamic ORd optimization) | 28878692 (corrigendum = 29230183) |
Tissue/organ methods and review
| Citation | PMID |
|---|---|
| Clayton RH, et al. Prog Biophys Mol Biol. 2011;104(1–3):22–48. (mono/bidomain review) | 20553746 |
Mitochondrial cardiac bioenergetics (Cortassa / O'Rourke)
| Citation | PMID |
|---|---|
| Cortassa S, et al. Biophys J. 2003;84(4):2734–2755. | 12668482 |
| Cortassa S, et al. Biophys J. 2006;91(4):1564–1589. (ECME) | 16679365 |
| Wei AC, et al. Biophys J. 2009;96(9):3510–3524. | 19289071 (search-confirmed) |
Curation notes / how to extend
- Use
model_type: KINETICfor ODE cell models;PHYSIOLOGICALorDIGITAL_TWINfor organ/whole-heart PDE and electromechanical models. There is no dedicated "PDE" enum value. - Map state-variable outputs to HP phenotypes via
ModelVariableDescriptorthresholds. Treat thresholds as model-calibrated illustrations, not clinical reference values (e.g. APD90 is the cellular correlate of the QT interval, not a direct QT measurement) and say so in the variable/mappingdescription, exactly as the CKD-MBD entry does. - All model evidence uses
evidence_source: COMPUTATIONALwith an exact-quote snippet from the model paper's abstract; validate with the standard reference and term validators before committing. - Next candidates in priority order:
RYR2_CPVT(Shannon–Bers/Grandi Ca²⁺-leak DAD models),Atrial_Fibrillation(Courtemanche),Brugada_Syndrome(reduced-I_Na tissue models),Hypertrophic_Cardiomyopathy(ToR-ORd, validated in HCM),Heart_Failure(Grandi remodeling).