FAQ

Why does the UI say "study" but the API says "experiment"?

They are the same object. Study is the user-facing label; experiment is the historical name in the database and REST API (/experiments/{id}).

I uploaded a file but analysis fails with "dataset not found"

Confirm ingestion status on Data → Datasets is ready, not failed. For CSV or 10x uploads, run Convert and wait for completion before starting Analyze.

Enrichment returns no terms

Check that your gene symbols match the reference database (HGNC for human demo data). Synthetic names like GENE001 will not match GO annotations. Re-run DE after fixing gene names in var.

Why are my results marked stale?

Changing metadata, design, contrasts, or re-running an upstream step (QC, normalization) invalidates downstream results. Open Runs → Stale outputs and re-run the listed steps.

Can I delete a pipeline run?

Yes, from Runs, unless the run is still active or referenced by a snapshot. Delete the snapshot first if needed.

Does the platform run R or Seurat?

No. Algorithms are Python ports inspired by Seurat, Bioconductor, and KnowSeq.

Where is bulk RNA-seq in the UI?

The bulk analysis API exists (POST /run/bulk) but the guided Analyze UI is not yet productized. Single-cell, spatial, and biomarker paths are fully wired in the app today.

Is AI interpretation available?

Not yet. Disease Evidence is available as a curated gene-association lookup, but AI-written summaries are not enabled. When enabled, AI will read structured results only — it will not compute p-values or fold changes.

How do I get help or report a bug?

This prototype is under active development. Contact your platform administrator for environment setup and support.