NMD-VCell · Virtual Cell Engine

Compose a disease-cell question. See what can actually be computed.

Specify the cell state, perturbation, time and endpoint. NMD-VCell returns the available evidence layer, the locked prediction layer and the experiment needed to connect them.

Compose a virtual-cell question Inspect capability levels

Current scientific boundary: evidence assembly and a limited same-assay baseline are available. Calibrated DMD perturbation response is not.

NMDVirtual
Cellevidence-aware

Cell state Perturbation Time Phenotype Uncertainty

Question composer

Define the biological transition before asking for an answer.

Local evidence check · no model inference

01Cell or state

02Disease context

03Perturbation

04Target

05Time

06Primary endpoint

Inputs assemble a research object in your browser. They do not call a disease-response model.

Current answer state Question assembled · prediction locked ZNF133 CRISPRi in DMD proliferating myoblasts can be linked to observed HepG2 response and an unperturbed human-myogenic reference, but not to a calibrated DMD state transition.

01Observed substrate21-gene HepG2 CRISPRi response matrix

→

02Biological context0 h human-myoblast reference

→

03Model taskSame-assay mean-response baseline only

→

04DMD truthNo independent matched outcome

MolecularLimited baselineSame-assay HepG2 only

Cell stateLockedNo conditioned DMD trajectory

FunctionLockedNo muscle functional outcome

ToxicityLockedNo matched safety readout

Cell–cellLockedNo causal cross-cell outcome

Decision-blocking gapIndependent DMD myogenic perturbation with a prespecified state endpoint

Open source evidence Design the missing test

Computable biological assets

Four layers connect the atlas to a future predictor.

Inspect all five datasets →

Layer 1 · Context

Human-myogenic reference

271 cells across 0, 24, 48 and 72 hours provide a measured, unperturbed differentiation axis.

Reference available

Layer 2 · Perturbation

Observed CRISPRi substrate

Twenty-one candidates have processed same-context HepG2 response vectors; nine have qualified human-myoblast screening context.

Context bounded

Layer 3 · Model

Frozen baseline task

Simple target-held-out baselines test recovery inside one assay. Cross-cell and disease transport are separate unevaluated tasks.

Executed · limited

Layer 4 · Truth

Prospective DMD outcomes

The registered study, prediction and outcome objects needed for calibration remain empty by design.

Not yet available

State transition contract

Measured reference and desired prediction are kept visually separate.

The current platform can locate a question along myogenic time. It cannot infer the counterfactual arrow from perturbation to DMD function until matched outcomes exist.

Open trajectory evidence Open the 24-slot pilot

Measured referenceMyoblast0 h

→

Measured referenceDifferentiate24–48 h

→

Context screenFusion9/21 candidates

⇢

Desired outputDMD function0 outcomes

Machine-readable product contract

The interface and its scientific ceiling are downloadable.

Use the schema to prepare a future virtual-cell run object. A valid object records modality, cell state, disease context, time, endpoint, evidence provenance and whether inference was executed.

{
  "cell_state": "proliferating_myoblast",
  "disease_context": "DMD",
  "perturbation": {"modality": "CRISPRi", "target": "ZNF133"},
  "time_hours": 48,
  "primary_endpoint": "cell_state_transition",
  "inference_status": "LOCKED_NO_DMD_OUTCOMES"
}

Open engine contract Open JSON Schema Download template