Large-scale single-cell mapping of concealed mitochondrial ageing
Jethwani, H.; Sukandar, K.; Green, A. P.; Insalata, F.; Golder, Z.; Schon, K.; Asquith, B.; Chinnery, P. F.; Jones, N. S.
Show abstract
Mitochondrial DNA (mtDNA) mutations have been linked to ageing. We find that the mtDNA mutations detected in routine bulk sequencing are greatly outnumbered by a class of concealed mutations that, despite being in a proliferative tissue, show little negative selection. Combining new scATAC-seq experiments with existing scRNA-seq, we map mtDNA mutations across 2.7 million cells from 311 individuals, covering 11 proliferative cell types and 5 tissues. We find that low-prevalence mutations (in[≤] 1% of assayed cells), concealed from routine bulk sequencing, accumulate markedly with age, with mutations unique to single cells (cryptic) predominating. We find a near universal accumulation rate across cell types and tissues, with, by age 50, approximately 40% of cells having at least one mutation affecting 30% of cellular mtDNA or more. At a correlative level we find that human age can be predicted from these concealed mutations but there is evidence of, possibly physiological, inter-individual variation. These low-prevalence mutations (unlike mutations tracking larger clones) show limited negative selection and coincide with gene-expression changes marking dysfunction.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Estimating the proportion of beneficial mutations that are not adaptive in mammals 95%
- R-loops and regulatory changes in chronologically ageing fission yeast cells drive non-random patterns of genome rearrangements 95%
- Reconstructing the history of founder events using genome-wide patterns of allele sharing across individuals 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.