Research Decision Queue

What should we test—and what would the result change?

NMD-VCell organizes open questions by the experiment or audit that can resolve them. It does not turn incomplete evidence into a target ranking.

See three open decisions Design an experiment

The unit of meaning

One disease question, one uncertainty, one action that can change our mind.

1Choose the cell state or function that matters.
2Separate direct observation from contextual evidence.
3Name the result that would support, downgrade or stop the hypothesis.

Myogenic state map

Start with the biological transition—not only the gene name.

Open measured trajectory →

01Proliferating myoblastMeasured reference0 h human-myogenic reference; no candidate DMD perturbation.

02Early differentiationMeasured reference24–48 h trajectory; descriptive, unperturbed context.

03FusionLimited perturbation context9 of 21 candidates assessed in a human-myoblast screen; not DMD replication.

04Maturing myotubeNot yet tested72 h reference exists; candidate perturbation function is unmeasured.

05Damage / stress responseNot yet testedFunctional DMD injury and recovery endpoints remain future study outcomes.

What this means: the current release can locate a question along a real human-myogenic reference, but it cannot simulate a DMD perturbation-conditioned fate.

Open decisions

Three different gaps require three different actions.

Ordered by decision type · not ranked

Decision 01DRAFT NOT REGISTERED

ZNF133 + MON1A myogenic test

Resolve whether ZNF133 or MON1A has a disease-context myogenic effect

Do ZNF133 or MON1A perturbations change a prespecified myogenic functional endpoint differently in DMD and matched-control cells?

What is already known

Observed CRISPRi response exists in HepG2 for ZNF133 (84 cells) and MON1A (94 cells).
Both targets were assessed in the GSE293514 human-myoblast fusion screen and neither met its FDR < 0.1 hit threshold.
Both current candidate records retain source-conflicted DMD direction and recommend a muscle-context assay.

What blocks the decisionNeither target has an independent perturbation repeated in a DMD muscle model.

Next discriminating action

Compare both targets in the same matched DMD/control human-myogenic assay, using qualified independent reagents and one frozen primary functional endpoint.

How would the result change the decision?

A reproducible DMD-specific effect would justify disease-context follow-up.
A similar effect in DMD and control would support a broader myogenic—not DMD-specific—interpretation.
A near-zero result inside a frozen equivalence margin would downgrade the hypothesis.
Viability loss, differentiation blockade or reagent discordance would classify the result as toxic or inconclusive.

Boundary: This card recommends a discriminating experiment. It does not predict that either target will improve DMD.

Compare ZNF133 and MON1A Open ZNF133 evidence Design the experiment

Decision 02PROVISIONAL HOLD IMPORT PENDING

GFOD2 evidence audit

Determine whether pending external evidence should change the GFOD2 record

Does the author-reported external assessment provide traceable quantitative evidence that survives the released import and audit rules?

What is already known

GFOD2 has an observed HepG2 CRISPRi response with 127 cells in the frozen aggregate.
GFOD2 was not included in the GSE293514 human-myoblast fusion library.
Its historical rank is preserved only for reproducibility; there is no current numeric rank.

What blocks the decisionThe reported external quantitative evidence, source provenance and analysis code have not been imported and audited.

Next discriminating action

Import the external result with source identifiers, quantitative values and analysis provenance, then run the released validation contract before changing status.

How would the result change the decision?

A complete, reproducible import can move the record from provisional hold to computational triage.
An incomplete or incompatible import leaves the record on hold.
Either outcome remains contextual evidence—not DMD perturbation efficacy.

Boundary: Audit resolution can change evidence status but cannot create a muscle, DMD or therapeutic validation claim.

Open GFOD2 evidence Read the evidence transition Download the import template

Decision 03DRAFT PANEL UNASSIGNED

24-slot pilot freeze

Freeze a balanced 24-slot DMD pilot before any outcome is visible

Which candidate, control and calibration conditions should occupy the 24 pilot slots under rules frozen before outcome access?

What is already known

The draft protocol defines 6 muscle-context, 4 high-uncertainty and 4 mechanism-diversity slots.
It also reserves 4 known controls, 4 random or negative controls and 2 toxicity controls.
No genes have been assigned, no study has been registered and no outcome has been measured.

What blocks the decisionSlot assignments, one primary endpoint, a minimally important effect and negligible-effect margin remain unfrozen.

Next discriminating action

Freeze the eligible universe, selection rules, slot assignments, controls, endpoint and decision thresholds in an immutable Study and Prediction package.

How would the result change the decision?

A frozen panel makes future hit-rate and enrichment claims auditable.
Controls allow calibration of false positives, nulls and toxicity.
Pre-outcome registration prevents retrospective relabelling of successes and failures.

Boundary: This is a prospective design decision. It contains no assigned panel, experiment, outcome or efficacy result.

Open the 24-slot protocol Review the no-wet-lab pack Open the empty registry

Progress that matters

The next milestone is a frozen decision—not another page.

Four of eleven no-new-wet-lab modules are executed or available. The 24 pilot slots are defined but unassigned. Study, prediction and outcome registries remain empty.

4/11evaluation modules executed or available

0/24pilot slots assigned

0prospective predictions registered

0outcomes returned

Freeze the pilot design Contribute evidence or outcomes