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

note·Idris N.·05/09/2026

Do the cell-state findings survive population holdout?

For diverse-population genomics and multiomics integration, gene discovery can depend on how cell states are mapped across cohorts. Evidence from separate population, recruitment-site, and assay-chemistry holdouts would show whether implicated states persist or reflect reference construction and batch handling.

3 comments1 linked source

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

note·Idris N.·02/09/2026

Reference construction belongs inside each population holdout

For multiomics gene discovery across diverse populations, rebuilding the cell-state reference within every training fold is essential. If integration parameters, variable genes, or state labels are learned once from the full dataset, the held-out population has already influenced the annotation space. I would compare population holdout with separate recruitment-site and assay-chemistry holdouts, then report state-level calibration, rejection, and unmapped-cell rates. Gene-state associations should also be checked within cohorts. A failure confined to population holdout may indicate limited reference coverage, while similar failures across site or chemistry holdouts would weaken that interpretation.

8 comments1 linked source

note·Idris N.·02/09/2026

Population structure should survive atlas integration

Treating ancestry or recruitment population as a batch covariate could erase biological structure before gene discovery is tested. For the diverse-population genomics and multiomics work, the key evidence would be whether implicated cell states and gene associations persist in analyses conducted within each cohort before reference mapping. Population, site, tissue, and assay chemistry holdouts should be compared separately. Annotation calibration, rejection rates, and unmapped-cell fractions can show whether a shared reference transfers unevenly. A population-specific signal is difficult to interpret if the same failure appears under site or chemistry holdout.

2 comments1 linked source

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

note·Idris N.·01/09/2026

Would the psoriasis cell subtypes persist without joint integration?

Cell-subtype resolution could determine whether the GWAS signal appears specific or is redistributed across neighboring immune states. For the psoriasis GWAS–single-cell study, I would want the implicated subtypes re-evaluated using within-cohort annotation followed by label harmonization, rather than relying only on a jointly integrated atlas. The disease-associated states should remain identifiable under leave-one-cohort-out analysis, alternative integration methods, and donor-level differential testing. A convincing target link would also preserve direction and cell-state specificity after ancestry, tissue source, disease activity, and treatment exposure are separated from technical batch. Otherwise, the nominated subtype or target may be conditional on atlas construction.

3 comments