What it proves
A released same-context perturbation substrate exists for governed candidate-level CRISPRi response auditing.
governed candidate-level aggregate · 2,000-dimensional response vector
Dataset Card · nmd-vcell-dataset-registry-record/1.0
Observed HepG2 CRISPRi candidate responses
A versioned, citable account of what was measured, which task this object may support and which prediction claims remain blocked.
01Sourceinternal-release-substrate
→02Typed contextHepG2 cell line
→03Qualified roleTrain and audit the same-context baseline
→04DMD prediction outcomeNot qualified
What it cannot prove
It cannot prove DMD rescue, muscle-cell function, patient response or disease-context transfer.
Dataset value is task-specific; no Dataset Card alone is evidence of candidate efficacy, patient response or clinical utility.
Best model use
Use as a permanent G0 same-assay baseline and leakage-control benchmark.
Train and audit the same-context baseline
Next experiment handoff
Pair the selected candidate with a matched DMD/control myogenic perturbation assay before any disease-response claim.
No matched muscle or DMD perturbation outcome
Identity & access
Source receipt
- Accession
- internal-release-substrate
- Licence
- Internal release licence pending institutional approval
- Source verification
- local governed object
- Stable page
/resource/dataset/nmdvcell-hepg2-crispri/
Biological context
What it represents
- Disease / context
- Context-limited; not DMD
- Species
- Homo sapiens
- Tissue / model
- HepG2 cell line
- Cell type
- HepG2
- Donor metadata
- not applicable cell line
Measurement design
What was measured
- Modality
- pooled perturbation response aggregate
- Observation unit
- governed candidate-level aggregate
- Size
- 21 governed candidate aggregates
- Count state
- component counts retained in governed candidate objects
- Endpoint
- 2,000-dimensional response vector
Perturbation design
Intervention contract
- Perturbation
- CRISPRi single-gene
- Dose
- CRISPRi single-gene; dose not modeled in released object
- Timepoints
- Frozen endpoint aggregate
- Direct DMD candidate outcome
- No
Replication & provenance
Receipt completeness
- Biological replicates
- limited release diagnostics only
- Technical replicates
- not exposed as public raw delta store
- Batch / platform
- processed release substrate; exact raw batch log not public
- Raw hashes
- raw delta store not public
- Processed hashes
- checksummed release objects available
- Analysis hashes
- derived tables in resource manifest
Benchmark partition
Permitted computational use
Direct same-context perturbation observation
G0_same_context_target_holdout
- Leakage grouping
- target holdout only no donor holdout
- Donor holdout
- No
- Disease-state holdout
- No
- Prospective holdout
- No
Readiness decision
Train and audit the same-context baseline
No matched muscle or DMD perturbation outcome
Associated resources