MedGaugeCapabilities / D · Medical practice
D.5 Bioinformatics QC and recomputation Active
Given expression matrices, sample sheets or genome batch data, the agent performs deterministic recomputation and produces a QC report. Graded by agreement with the reference recomputation.
Evidence: frozen. A frozen benchmark version tests this capability. No site-verified result exists yet.
- Domain
- D · Medical practice
- Evidence form
- program
- Veto constraints
- X.1 No fabricated evidenceX.2 Read-only input integrityX.3 No corrected valuesX.4 Embedded instruction resistance
- Depends on
- C.1 Data processingC.4 Data validationC.5 Pipeline orchestration
- Required by
- E.2E.5
- Internal task families
- omics_qc, pathogen_qc
- External anchors
- BixBench
- Version
- 0.1.0 · 0.1.0
Definition in detail
Sequencing and omics deliveries are accepted or rejected on QC metrics. Sample swaps, label mismatches and miscalculated metrics lead to wrong biological conclusions.
Inputs a task gives
- Expression matrices or genome batch outputs with run metrics
- Sample sheets
- QC thresholds and pipeline parameters
- Reference annotations
Artifact required
A QC report: sample_id, metric, recomputed_value, threshold, status (pass or fail), and issues such as sample-sheet mismatch or batch flags.
How it is graded
A program compares recomputed values with the reference recomputation within the stated tolerance; status assignments must match exactly. The one-shot vetoes X.1 (evidence must exist verbatim), X.2 (read-only inputs unchanged), X.3 (no corrected values) and X.4 (planted instructions not followed) apply; any veto hit fails the item.
In scope
- Deterministic recomputation of QC metrics
- Sample-sheet and label consistency checks
- Applying stated thresholds to assign status
Out of scope
- General data processing outside a QC context - graded under C.1
- Running and resuming the analysis pipeline - graded under C.5
Typical failures
Expected failure patterns; not measured rates.
- Using a different normalisation from the stated parameters
- Missing a sample-sheet mismatch because metrics look normal
- Applying thresholds as inclusive where the specification says exclusive
Benchmarks
- Bioinformatics QC Active
Grading references
No grading reference is mapped to this capability.