Plan an experiment
Study design
Turn one candidate decision into a predeclared experiment.
Search a gene. For the 21 governed candidates, the planner loads the released decision contract, prefills the question and assay route, and requires human confirmation before generating a draft. It does not rank candidates or upgrade evidence levels.
Step 1 of 5
Selected gene: DNM1. Current evidence level: L2 (first arm). Blocking gap: GAP-07. Recommended route: Computational replication.
Recommended examples: ZNF133 · CPEB1 · GFOD2.
Module meaning contract
Design · what this module can change
Question answeredWhat falsifiable experiment would close the named evidence gap and what result would change the decision?
Scope & evidence unitVersioned Study Cards currently cover the 21 governed DMD candidates. Other disease modules provide bounded next-experiment recommendations, not registered Study Cards.One versioned Study Card with estimand, endpoint, replication, QC, analysis and stopping rule.
ProducesA downloadable draft design object with explicit registration gaps and links to its evidence inputs.
Cannot establishA draft is not a registered study, executed experiment, measured outcome, target validation or efficacy result.
Next handoffInspect lifecycle state in Registry; registration and execution remain prospective. Continue →
Product contract · WORKFLOW-OUTCOME-CARD-1.0
Workflow Card: What experiment can change the conclusion?
EstimandContext · dose · timeBiological unitControls + endpointsQC + decision ruleOutcome return
Candidate decision contract · governed prefill
Candidate must answer one bounded question before a draft is created.
The fields below come from the released Candidate Decision Registry. Review them before accepting the automatic experiment prefill.
Review result-to-decision rules
SupportiveNot loaded
NullNot loaded
Inconclusive / QC failureNot loaded
Source: Candidate Decision Registry. Confirmation records design intent; it does not register or execute an experiment.
Five-step experiment planner
Answer four more questions to create a Study Card.
What this means: candidate-specific fields are prefilled from the released decision contract. You can refine the cell model and endpoint, but the exported draft preserves the original contract and its result rules.
Planner visual builder
Turn one gene into a bounded experiment route.
The planner is a preregistration-style experimental workflow: it resolves a gene, shows the frozen evidence state, exposes the blocking gap, selects the current assay route and lists the registration fields still open.
BoundaryThis builder does not rank candidates, change evidence levels, compute target scores, infer therapeutic direction or make disease-prediction claims.
Draft Study Cards21versioned Study Cards
Action classes5current next-step categories
L3b replications0same-perturbation evidence absent
Prediction statusP0P1/P2 remain locked
1Resolve geneHGNC identifier and typed record
2Evidence stateL1/L2/L3a fields only
3Blocking gapgap code stays explicit
4Assay routeaction-specific scaffold
5Register draftrequires frozen fields
Current Study Card action map
Test in human myogenic cells2 Study Cards
Run an independent human myogenic perturbation experiment.
Confirm the gene is detectable4 Study Cards
Confirm target detectability before perturbation escalation.
Review safety and biological context7 Study Cards
Resolve dependency, toxicity or context risks before escalation.
Reproduce the result independently7 Study Cards
Repeat the source-state analysis in an independent frozen workflow.
Pause until external evidence is audited1 Study Card
Import and audit the reported external evidence before changing state.
Shared planner link restored on the server
Selected gene: DNM1
HGNCHGNC:2972
Current evidence levelL2
Blocking gapGAP-07Independent computational replication is absent
Recommended routeComputational replication
Comparison object
comparison-not-availablePlanner object
PLN:v1.2.0-measured-dmd-evidence:DNM1:1.0-DRAFTStudy Card object
SC:v1.2.0-measured-dmd-evidence:DNM1:1.5-DRAFTOutcome slot
OUT:v1.2.0-measured-dmd-evidence:DNM1:PENDINGSelected gene: DNM1. Current evidence level: L2 (first arm). Blocking gap: GAP-07. Recommended route: Computational replication. Object trace: comparison-not-available → PLN:v1.2.0-measured-dmd-evidence:DNM1:1.0-DRAFT → SC:v1.2.0-measured-dmd-evidence:DNM1:1.5-DRAFT → OUT:v1.2.0-measured-dmd-evidence:DNM1:PENDING.
Candidate decision contract restored on the server
DNM1 · review before creating a draft
Result-to-decision rules
Concordant / feasibleIf direction and uncertainty reproduce under the frozen plan, retain the candidate and move to its next context-specific gate.
Equivalent null / stopIf the signal does not reproduce despite adequate coverage and QC, retain the negative result and do not escalate from the original analysis alone.
Not evaluable / QC failureIf coverage, mapping or source design prevents a fair comparison, mark the replication non-evaluable rather than positive or negative.
Confirmation required: the interactive planner prefills these fields but will not create a candidate draft until the researcher confirms this contract.
Stable governed records
Versioned Study Cards grouped by current action
All 21 routes are static, shareable and tied to v1.2.0-measured-dmd-evidence, schema 1.1 and build EA-20260817-57. JSON, YAML and TSV exports are available on every card.
Test in human myogenic cells
2 Study Cards
Run an independent human myogenic perturbation experiment.
Confirm the gene is detectable
4 Study Cards
Confirm target detectability before perturbation escalation.
Review safety and biological context
7 Study Cards
Resolve dependency, toxicity or context risks before escalation.
Reproduce the result independently
7 Study Cards
Repeat the source-state analysis in an independent frozen workflow.
Pause until external evidence is audited
1 Study Card
Import and audit the reported external evidence before changing state.