v1.0.0-database-resourceSchema 1.1Model ridge-safe-v2.3Benchmark repeated-fold-v2.2Build EA-20260729-15v1.0 is a database and evidence-governance release; it is not a validated disease-prediction or clinical decision-support release.
Current selection: unresolved requires registration
State whether the disease-associated direction is hypothesized as causal, compensatory or accompanying, and preserve the opposite-direction alternative. DMD direction and counteralignment never choose an intervention automatically.
Allowed registered hypotheses: activation; inhibition; bidirectional exploration; direction not identifiable.
Comparator and controls
Frozen null and simple baselines.
Leave-one-source-out analysis.
Negative-control genes or permuted labels.
Complete unmatched-identifier report.
Secondary endpoints
Pathway concordance.
Leave-one-source-out robustness.
Coverage and null-result diagnostics.
State-transition extension
Mechanism hypothesis and registered time axis
Scientific object: perturbation × cell state × disease context × time × phenotype
Required before registration: state the proposed early molecular mediator, the expected cell-state transition and the downstream functional consequence.
6–12 hearly molecular or signalling response
planning default requires assay calibration
24–48 hregulatory program and cell-state transition
planning default requires assay calibration
4–7 ddifferentiation and functional phenotype
planning default requires assay calibration
Cell context fields
Required before registration: healthy, DMD or isogenic corrected. Required before registration: proliferating myoblast, early differentiation, fusion or maturing myotube. Required before registration; preserve donor-specific estimates.
Cell–cell consequence
Current status: not assessed. Future levels: conditioned medium; two-cell co-culture; three-dimensional muscle model; spatial perturbation model.
Blocking factors
Study/donor.
Batch.
Cell state.
Dataset-specific covariates.
Assay QC thresholds
Minimum donor/sample coverage.
Gene detectability.
Frozen mapping and exclusion thresholds.
Minimally important effect
Required before registration; derive from assay biology or a justified pilot and store the numeric value with units.
Negligible-effect margin
Required before interpreting a null result; store a symmetric or asymmetric numeric margin with units.
Multiplicity and missing data
Frozen gene and pathway families with declared adjustment method.
Define exclusions before unblinding; report all missing units and reasons; do not single-impute primary outcomes without a prespecified sensitivity analysis.
Replication rules
Not applicable unless the independent dataset contains multiple perturbation reagents.
A result must persist beyond a single donor/sample and disclose leave-one-unit-out sensitivity.
Stop rules
Stop or classify as infeasible if a required numeric QC threshold fails.
Do not interpret a nonsignificant result as no material effect without a negligible-effect interval.
Do not change canonical candidate status automatically; require governed review.
Time and cost1–4 weeks after data access; planning estimate only. Institution- and assay-dependent; obtain a local itemised quote before registration.
Preregistration statusdraft requires direction rationale numeric effect margin sample size and qc thresholds
Lifecycle ruleThis draft is editable. Registration requires a timestamp and checksum; a registered revision is immutable and any correction must supersede it.
Escalation ruleEscalate from L2 to L3b only after independent context-matched perturbation replication; L4 requires replicated DMD-relevant muscle evidence.
Evidence transitionCurrent evidence state → predeclared independent test → governed evidence-level review.
Data-release planRelease the frozen card, protocol identifiers, analysis code, complete denominators and results irrespective of direction; never overwrite the registered card.
Boundary: This Study Card is an evidence-gated design scaffold, not a protocol, power calculation, safety claim, prediction or therapeutic recommendation.