Life sciencesResearch LabsStandard · methods + data uploadedGuardian · academic integrity1d ago · Genome Engineering Lab

"Is the CRISPR off-target editing rate reported in this foundational study reproducible, and what does the replication literature show?"

Principal investigator · Pre-submission reproducibility check · Methods and detection data submitted as context

Illustrative session · augle.com
Finding · Phase 3 synthesisContested
41%
Ensemble confidence
CONFIDENCE GRADE
Contested

Convergence exists but shares one detection method across all studies. Method-independent reproducibility: Contested.

Calibration basis
Confidence grade

Finding is the confidence grade itself. Confidence grade is the finding. For an open research question, the grade — not a numeric score — is the complete result.

The ensemble returns a Contested finding on the reproducibility of the reported off-target editing rate. Multiple studies report broadly consistent off-target rates, which superficially looks like robust replication. But the Contrarian's Strong objection is unresolved: nearly all of these studies rely on the same detection assay, so their agreement may reflect a shared methodological blind spot rather than independent confirmation. Convergence without methodological variation is not the same as reproducibility. The within-assay claim (results replicate when the same detection method is used) is Probable; the method-independent claim (the true off-target rate is as reported, regardless of assay) is Contested. The actionable output: confirm with an orthogonal detection method before treating the rate as established.

Confidence grid · 5 agents × 20 units
Agreement / confidentShorter row = lower confidence / dissentUnfilled
Reopen conditions · evidence-triggered, not market-triggered
RC1
Independent replication using an orthogonal off-target detection assay (e.g. a method mechanistically distinct from the one shared across current studies)
Trigger: Orthogonal-assay replication · Direction: Established grade possible if the rate holds
RC2
A head-to-head comparison of detection assays on the same edited samples, quantifying method-specific bias
Trigger: Assay-comparison study · Direction: Resolves the shared-method objection
RC3
Reframing the manuscript claim from "the off-target rate is X" to "the off-target rate is X as measured by [assay], pending orthogonal confirmation"
Trigger: Claim revision · Direction: Probable grade for the assay-scoped claim
Phase transcript · all three rounds
Cartographer
Gemini 3.1 Pro · T=0.8
61%

Settled ground: CRISPR-Cas9 produces off-target edits at some non-zero rate. Several independent groups report off-target rates in a broadly similar range for the guides in question. Detection sensitivity depends heavily on the assay used.

Contested terrain: Whether the reported rate is a property of the editing or of the detection method. Whether cross-study agreement constitutes independent replication given shared methodology.

Unknown territory: How the reported rate behaves under a mechanistically orthogonal detection assay. Whether the shared assay systematically under- or over-counts a class of off-target events.

Knowledge gaps entered: (1) Orthogonal-assay replication — not present in the uploaded literature. (2) Head-to-head assay-bias quantification — does not exist for these samples.

Methodologist
GPT-4o · T=0.5
54%

Methodological-independence concern flagged: The replication set is not methodologically independent — the studies share a common detection assay. Agreement among non-independent measurements does not raise confidence in the same way independent replication does. The apparent robustness may be an artefact of shared method bias.

Evidence ceiling: the method-independent reproducibility claim is capped at Contested until an orthogonal assay confirms the rate. The within-assay reproducibility claim (same method, same result) is Probable. No node reaches Established without method-independent confirmation.

Contrarian
Claude Sonnet 4.6 · T=1.0
35%

@Cartographer — Steelman: Multiple groups reporting a consistent rate is genuine evidence, and the guides and conditions are well-characterised. A working biologist would reasonably treat this as reproducible.

Strong objection [Phase 1]: "The replication attempts use a single detection assay across all studies. Convergence without independent methodological variation is not reproducibility — it is the same measurement repeated. If the shared assay has a systematic blind spot for a class of off-target events, every study inherits it and they will agree while all being wrong in the same direction." Resolution condition: Replicate the rate with a mechanistically orthogonal detection assay on the same samples.

Dissent register · all Contrarian objections1 Strong Unresolved · 1 Moderate
Strong@Contrarian → @CartographerPhase 1 · carried to Phase 3Unresolved
Steelman

Consistency across independent groups is genuine evidence, and the guides and editing conditions are well-characterised. Treating the rate as a working value is reasonable.

"The replication set shares a single detection assay. Convergence without independent methodological variation is repetition, not reproducibility — a shared blind spot would make every study agree while all being wrong in the same direction."

Resolution condition

Replicate the rate with a mechanistically orthogonal detection assay on the same samples

Moderate@Contrarian → @MethodologistPhase 2Actionable
Steelman

Single-assay replication is common in this subfield because orthogonal assays are costly and low-throughput. The convergence is still reassuring within its limits.

"The manuscript states the off-target rate as a property of the edit rather than of the measurement. Even on the within-assay reading, the claim should be scoped to the assay used."

Resolution condition

Scope the reported rate to the detection method in the manuscript text

Pragmatist action item: Scope the rate to the assay and commission an orthogonal-assay replication before submission

Guardian integrity log · academic integrity mode96% · Clean
96%
Overall integrity score
Academic integrity mode

No retracted papers. One preprint down-weighted below the primary-claim threshold. Methodological-independence gap flagged across the replication set (shared detection assay). Self-citation ratio within field norms. Phase boundary clearances issued at P1/2 and P2/3.

Source quality
97%
Retraction check
100%
Preprint flag
94%
Self-citation ratio
96%
SVS verification log · 13 citations checked
Foundational study — reported off-target editing rate for the guide set
Peer-reviewed · Verified
Replication A — off-target rate, same detection assay
Peer-reviewed · Verified
Replication B — off-target rate, same detection assay
Peer-reviewed · Verified
Preprint — additional replication, same assay family
SVS_PREPRINT — bioRxiv preprint, not peer-reviewed · Evidence node capped at Probable · Flagged for substitution
Preprint · Flagged
Methods review — comparison of off-target detection assays
Peer-reviewed · Verified
+8 additional citations verified · 0 retracted · 1 preprint flagged above
Session metadata
Session IDlife-crispr-reproducibility
VerticalResearch Labs
DepthStandard
Runtime17m 05s
Attachmentmethods + data
Guardian modeAcademic
Guardian score96% · Clean
Citations11 · 1 preprint
Dissent flags1 Strong · 1 Mod
Calibration status
Calibration basis
Confidence gradeConfidence grade is the finding

For open research questions, the confidence grade is the complete finding — calibrated against the evidence base, not against a binary resolution outcome.

Finding gradeContested · 41%
Agent confidence
Cartographer
Gemini 3.1 Pro
52%
Methodologist
GPT-4o
48%
Contrarian
Claude Sonnet 4.6
28%
Synthesizer
GPT-4o
44%
Pragmatist
Grok 4.1 Fast
42%
Export & share
Illustrative session · augle.com