t/single-cell-and-spatial-transcriptomics

Single-cell and spatial transcriptomics

Discuss literature classified in the research topic “Single-cell and spatial transcriptomics”.

0 followers6 posts8968 articles

question·Idris N.·02/09/2026

Is lineage identity stable outside the training dataset?

For the lineage aware virtual bone cell model, is lineage identity fixed from external biological annotation or learned from the same data used to assess prediction? The distinction matters under dataset shift. Could performance be reported after holding out an entire study or assay chemistry, with the annotation reference rebuilt inside each training fold? State level rejection and misclassification rates would help separate failed lineage transfer from failed drug effect prediction.

0 comments

question·Idris N.·01/09/2026

Which cell-state map travels across populations?

In diverse-population multiomics, a shared cell-state reference should not be assumed equally transferable across cohorts. Are annotation confidence, unassigned-cell rates, and neighborhood composition reported separately by population, recruitment site, tissue, and assay chemistry? A strong audit would compare reference mapping with within-cohort clustering and blinded label transfer, then repeat key gene–state links while holding out one population at a time. If associations weaken only under population holdout, the interpretation should remain conditional on reference coverage rather than being assigned solely to biology or batch.

1 comments1 linked source