Post-ONT CpG classification — named libraries and POD5

Actual UI, pointer drags, temporary configuration saves and checks.
Synthetic reads/resources only; no analysis, downloads or patient data.

One named folder per ONT library
Each ONCO-style library has fastq_pass and pod5_pass. This demo restores three libraries from their saved ONT project.

Matching POD5 is selected
Each library keeps its own passed signal reads. Compatible Dorado and classifier models are required; these are synthetic fixtures.

1 · Drag samples
Synthetic example: ONCO-001 → MARLIN; ONCO-002 → Sturgeon. For real data use clinical records; never infer diagnosis from names.

Unknown is separate from Cancer or Normal
Cancer/Normal records the clinical role. These columns only choose a CpG classifier. Return unsuitable samples to Unknown / do not classify.

2 · Save the chosen group
Choose MARLIN and save. Its new project contains only ONCO-001.

2 · Save the chosen group
Choose Sturgeon and save ONCO-002 separately. The other group and Unknown sample are excluded.

Confirm your Sturgeon license
Only acknowledge this after obtaining Sturgeon and accepting its terms. This demonstration uses synthetic resources.

3 · Run only when ready
Use valid POD5 and compatible models, resolve checks and license requirements, then click Run explicitly. This demo stops after saving.