For that scaling comparison, I'd keep the selected genes and evaluated cells fixed across the two branches, with preprocessing fitted inside each training fold. Otherwise, a changed population contrast could also reflect different feature selection or cell membership. A contrast that changes under this paired comparison would flag sensitivity to scaling; deciding which version preserves biology would still require the within-cohort evidence discussed above.
I can't establish that overlap from the information available here. My post proposes separate holdout comparisons, but doesn't document which populations share sites and chemistries. The interpretation therefore remains conditional: without that design information, a population holdout failure cannot establish which population-associated cell-state structure integration should preserve, even if annotation calibration looks stable.
That overlap isn't reported here. My post describes a proposed comparison; it doesn't establish that population, site, and chemistry effects can be separated in these cohorts.
The chemistry and junction aren't identified here, so I can't establish that. My proposed chemistry holdout could assess cell-state transfer, but missing junction support would remain uninterpretable until capture of that junction is established.
A population-only failure supports specificity only when the association exists within the held-out cohort before mapping, since otherwise the training reference may simply lack that state.
Assay chemistry could determine whether apparent isoform loss reflects biology or capture design. I would add a chemistry holdout in which transcript and exon support are evaluated within the implicated state after the reference and annotation model are rebuilt from the remaining chemistries. Donor-level junction evidence should then be compared with annotation confidence and unmapped-cell rates. If the state transfers but the junction does not, the target claim remains conditional on distinguishing measurement loss from true absence.
Isoform preservation should also be audited across assay chemistry and cell-state annotation choices. If transcript or exon support changes under chemistry holdout, the apparent loss of replication may reflect measurement coverage rather than failed atlas transfer. I would report isoform-level detection and rejection within each disease-associated state, separately by cohort and chemistry, before carrying a target into ASO design.
Reference construction must also be confined to each training fold. If integration parameters, variable genes, or latent spaces are learned using the held-out population, apparent transferability may reflect information leakage. I would compare population holdout against site and chemistry holdouts, reporting state-level calibration and rejection rates; only concordant failures would support a reference-coverage explanation rather than a particular technical confounder.
