Appendix 2—table 1. Seven single-cell RNA sequencing studies of different tissues where age-related increase in transcriptional noise was measured.

The number of cells (N. cells) in the table is the size of the dataset prior to quality control. The Noise column states whether an increase in transcriptional noise was reported in some/all cell types in the original articles. The Scope column summarizes the cell types where age-related increase in transcriptional noise was reported. The Method column specifies how transcriptional noise was measured in the original articles.

Dataset	Tissue	Organism	N. cells	Noise	Scope	Method
Enge et al., 2017	Pancreas	Human	2544	Yes	In Beta cells.	(1) Biological over technicalvariation, (2) wholetranscriptome-based Euclidean distance to cell typemean, (3) invariant gene-based Euclidean distance to celltype mean.
Martinez-Jimenez et al., 2019	CD4+ T cells	Mouse	1513	Yes	Single cell type studied.	Percentage of cells expressingthe core activation program.
Angelidis et al., 2019	Lung	Mouse	14,813	Yes	In most cell types.	Distance to cell type mean.
Kimmel et al., 2019	Lung, spleen, kidney	Mouse	30,255 30,512 29,815	Yes	In many cell types.	(1) Overdispersion of genes, (2) invariant gene-based Euclidean distance to cell type mean, (3) whole transcriptome--based Manhattan distance to cell type mean.
Ximerakis et al., 2019	Brain	Mouse	37,069	No	Differences in magnitude and directionality between cell types.	Coefficient of variation of (1) all genes, (2) mitochondrial genes, (3) ribosomal genes.
Salzer et al., 2018	Dermal fibroblasts	Mouse	731	Yes	Single cell type studied.	Compactness of clusters on PCA plot.
Solé-Boldo et al., 2020	Skin	Human	22,142	Yes	In dermal fibroblasts.	Less clear GO (Gene Ontology) annotations.