Spike-in-normalised single-cell RNA-seq reveals cell-type-specific transcriptional repression during ageing

Transcriptome analyses are widely used for biomarker discovery and to gain insights into normal processes and diseases. Age-related changes in gene expression inferred from RNA-seq are typically reported relative to the transcriptome composition using library-size normalisation. As such, absolute changes in transcript abundance with age remain poorly characterised. Here, using external spike-in normalisation in the Tabula Muris Senis dataset, we quantify age-related variation in total mRNA content and gene expression across mouse cell types. We observe widespread changes in total mRNA abundance, with decreases predominantly in non-immune cell types and increases predominantly in immune cell types. In parallel, the number of genes expressed declines across most cell types, including immune populations. Differential expression analysis based on spike-in-normalised counts identifies genes consistently downregulated across cell types, enriched for functions in RNA metabolism and protein processing. Furthermore, genes downregulated during ageing and during proliferation arrest show partial overlap, suggesting that these transcriptional changes may share regulatory processes. Together, these results are consistent with a general repression of transcriptional and metabolic activity with age, modulated by immune-specific responses. More broadly, our results demonstrate that conclusions drawn from transcriptomic ageing studies can depend strongly on whether gene expression is interpreted in relative or absolute terms, highlighting the importance of absolute normalisation approaches for the analysis of age-related transcriptomic change