Barcode failure fixture: unsafe one-edit correction

Category: Report zoo
Assay type: inline_barcode
Evidence level: Small fixture
Dataset: Committed barcode failure-mode FASTQ fixture
Condition shown: Observed barcode sequences are within one edit of multiple targets, so k=1 rescue is unsafe.
Primary report: examples/barcode_autopsy/failure_modes/README.md

What This Shows

DotMatch can show why one-edit barcode rescue can corrupt sample assignment.
The same observed barcode can be retained as ambiguous instead of assigned to the first compatible sample.

What It Does Not Show

This is a diagnostic fixture, not a public demultiplexing benchmark.
Collision-safe barcode design remains the user’s responsibility.

Comparator And Validation

Comparator semantics: The fixture checks deterministic barcode collision and ambiguity classifications under fixed-window assignment semantics.

Validation: The expected findings table includes duplicate barcode, unsafe one-edit collision, and ambiguous-read rows.

Report Examples

Expected unsafe-correction findings (tsv)

Raw Artifacts

Reproduce Or Check

make barcode-validation-ready