Profile a metagenomic sample into an auditable research report

Why use this

1. A sequencing analysis report can look comprehensive while concealing mismatched database versions, omitted stages or assumptions about what a detected marker means. A table of assigned organisms is not, by itself, evidence for a clinical or causal conclusion.

2. A research-provenance packet could record the intended input handling, tool roles and interpretation boundaries before analysis begins. Domain reviewers could inspect whether the proposed outputs answer the research question without treating workflow completeness as biological validation.

Why an agent / simpler alternative

Useful judgment

Organize analysis-stage dependencies and flag provenance gaps or interpretations that exceed the proposed evidence.

Simpler alternative

Use an established institutional analysis service when its validated scope already fits the research question.

Unsuitable when

The workflow would autonomously diagnose, recommend treatment or process unapproved identifiable human data.

Proposed inputs

1. Name

   Research scope

   Owner

   Domain investigator

   Minimum scope

   Non-clinical question, approved output categories and optional-stage decisions

   Quality check

   Identify interpretation limits and excluded analyses.

2. Name

   Analysis metadata sample

   Owner

   Authorized data steward

   Minimum scope

   Fictional sample descriptors, tool/database versions and provenance fields

   Quality check

   Flag missing consent, governance or version information.

Access and setup — proposed or unknown

Proposed minimum

1. Text-only fictional metadata and an analysis-provenance plan without sequence data.

Compatibility unknowns

1. Unknown: current runtime, reference databases, tool compatibility and computational requirements.

External actions requiring approval

1. Real input processing, expert-approved parameters, compute budget, data sharing and scientific publication.

Expected deliverable and review checks

Proposed artifact

Research output specification with provenance fields, stage exclusions and expert interpretation questions.

Critical checks

1. Each intended result is linked to the relevant analysis stage and database version.

2. Marker and abundance outputs are not presented as diagnoses or treatment guidance.

Missing output evidence

No new workflow run or output inspection was performed for this overlay; use the canonical source-outcome references.

Failure signals and proposed responses

1. Signal

   A database version is missing from the proposed results manifest.

   Why it matters

   Later comparisons may be difficult to interpret or reproduce.

   Proposed response

   Require the provenance field before accepting the reporting plan.

2. Signal

   An omitted optional stage is described as yielding negative findings.

   Why it matters

   Not analyzed is being confused with absence.

   Proposed response

   Label the stage not performed and remove unsupported result language.

Effort and cost — unknown

Setup work

1. Define data governance and result-provenance requirements.

Ongoing review

1. Inspect interpretation boundaries and later expert-reviewed outputs.

Unknowns

1. Unknown: compute and storage costs, database requirements and analytical validity.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

A fictional analysis manifest omits a database version and labels an unperformed optional stage negative.

Acceptance checks

1. Flag both provenance and status errors.

2. Keep all biological findings unspecified.

Stop conditions

1. Sequence data, analysis commands, clinical interpretation or database retrieval is required.

Limits

1. This document check supplies no taxonomic, resistance, pathway or clinical validation and is not an executed research analysis.