Evaluating a cross-platform kidney atlas as a reusable reference.
Scilytica is studying whether a cross-platform kidney atlas can serve as a stable reference for new CosMx and Xenium patient samples.
The design keeps CosMx and Xenium measurement models separate, uses a shared single-nucleus RNA-seq reference and label ontology, and assigns complete patients to reference, calibration, or held-out roles. Candidate configurations are selected without using the held-out patients.
Clinical samples may arrive gradually over months or years, while research must continue before the full cohort is available. Rebuilding the atlas or changing how each new sample is processed, mapped, reviewed, and reported can turn the analysis into a moving target. Differences between patients may then reflect changes in the analytical process as well as changes in the underlying biology.
The primary question is whether a frozen mapping strategy can map previously unseen patients without changing reference-owned state and outperform a qualified frozen-reference baseline. Software, provenance, and workflow controls support that comparison by preserving the methods, parameters, quality controls, and evidence associated with each result.
Study design and validation are in progress. Results will be shared after review.