Cells of the same chronological age do not necessarily show the same aging state. On September 12, 2026, npj Aging made the scMLEAge study available. It estimates age from individual cells’ transcriptional data, bringing differences within organs and between cell types into the analysis. The paper is currently an accepted, citable advance-publication version.
From a tissue average to individual cells
The method proposed by Chanyue Hu and Matteo Pellegrini uses single-cell transcriptomic counts within a Bayesian statistical framework to estimate the age most consistent with those data. The authors applied the model to the Tabula Muris Senis mouse aging atlas to analyze organ- and cell-type-related age features, reporting better predictive performance than the conventional regression methods compared. Paper page
A transcriptome records gene activity in a cell at a given time. Analyzing mixed cells gives an average picture of a tissue, whereas single-cell analysis can identify different states beneath that average. This heterogeneity in aging is scMLEAge’s point of entry.
Molecular age still needs a functional interpretation
The model’s output reflects the relationship between molecular features and ages in its training data. A predicted age that differs from chronological age may relate to cell state, but it may also be affected by the data and the model’s scope of applicability.
TASCAP follows this work because it offers a finer observational scale for studying differences in aging. The next questions are whether the differences identified by the model correspond to independently measured cellular functions, and whether they can be reproduced in other datasets.
The current research is based on mouse data. It advances computational analysis of single-cell aging; use in human health assessment would require relevant human data and validation for the intended purpose.