scCody distinguishes aging-aligned from divergent disease-associated cellular remodeling

scCody separates disease-associated cellular remodeling into aging-aligned and orthogonal components, revealing distinct patterns in Alzheimer’s disease versus COVID-19.

• Why it matters: Aging clocks compress molecular change into one estimate, leaving disease effects that track healthy aging indistinguishable from those that diverge.
• What they did: scCody is a generative model trained on cross-sectional single-cell cohorts using population-level optimal transport plus endpoint-consistency regularization to reconstruct unseen ages and infer donor age without age-regression training.
• The result: In Alzheimer’s disease, neuronal populations show substantial aging-aligned remodeling with shared orthogonal pathology-linked changes, while COVID-19 shows weaker alignment and stronger orthogonal immune remodeling, including interferon-linked symptom-duration effects in CD16 monocytes.

AI summary of the abstract · may contain errors · not medical advice · read the paper.