Dataset Processing Table

One row per dataset: ingestion track, QC, normalisation, batch-correction level and batch key, annotation source and reference, and concordance with original author annotations. Methodological background is documented on the Data Processing page. Badges with dashed borders carry the dataset’s processing note — hover to read it.

manifest generated: 2026-09-29
DataID Species Tissue Input track Cells (post-QC) Annotation Batch correction Detail

“Undocumented” is shown explicitly wherever a processing attribute has not yet been back-filled from the pipeline records — it is a transparency statement, not a missing-value error. Rows without a DataID are atlas bundles deployed on the site but not yet registered in datainfo.csv. By default the best-documented datasets are listed first — click any column header to re-sort, or use Detail to view a record’s full information.