Cell-cycle analysis: model checks and public-data preflight

What was checked

• DNA-content mixture fitting on balanced, G1-enriched, S-enriched, and G2/M-enriched synthetic profiles with known expected distributions.
• FCS parsing and cell-cycle fitting on bundled synthetic DAPI and PI fixtures.
• Preflight selection on public FlowRepository accessions FR-FCM-Z55E and FR-FCM-Z55H, including morphology, singlet, and PI-signal checks before reference comparison.

What the result means

The synthetic fixtures test whether the implemented models recover their known inputs. The public-data run labels samples as candidates, review-fit cases, background controls, or insufficient-event cases; these are triage outcomes, not proof that every biological sample is correctly classified. Several samples require manual review.

The public datasets are associated with supplementary figures in the Sam68 study. FlowApp's use here is limited to software preflight and does not independently reproduce or validate the paper's biological conclusions.

Reproduce the checks

In the repository, run npm run validation:cell-cycle:model, npm run validation:cell-cycle:fcs, and npm run validation:cell-cycle:flowrepository. The preflight expects the public FlowRepository files to be present in the local validation cache.

See the FR-FCM-Z55E dataset, FR-FCM-Z55H dataset, and the associated Sam68 paper.