Back

Somatic mutations and single-nucleus transcriptomics reveal uniquely human properties of neuronal aging

Caglayan, E.; Lamba, I.; Manam, M. D.; Finander, B.; Luquette, L. J.; Exposito-Alonso, D.; Zhao, S.; Jin, B.; Miller, M. B.; Park, P. J.; Sherwood, C. C.; Walsh, C. A.

2026-09-24 neuroscience
10.64898/2026.09.22.753624 bioRxiv
Show abstract

Human brain aging is frequently studied in model mammals that rarely show the natural age-related neurodegenerative conditions that afflict humans, yet little is known about how genomic stability compares across species during aging. Although somatic single nucleotide variant (SNV) and short insertion/deletion (Indel) mutations accumulate at rates that inversely scale to lifespan in colon cells - leaving diverse mammals with similar end-of-life burdens of mutations - here we show that cerebral cortical neurons accumulate SNVs at annual rates that are remarkably similar across six mammalian species (human, chimpanzee, rhesus macaque, marmoset, ferret, mouse) resulting in >12-fold more mutations in human neurons at the end of life compared to mouse neurons. Despite the conservation of overall annual mutation accumulation rates, mutational patterns--and hence likely mutagenic mechanisms--show sharp differences between species, with a nucleotide substitution pattern linked to neurodegenerative diseases accumulating almost exclusively in humans during aging. Coinciding with their higher mutational burden, single-nucleus transcriptomic analyses reveal pervasive age-associated proteostasis and mitochondrial dysregulation in human neurons that is far less pronounced in aged chimpanzees and rhesus macaques. These findings show that the pervasive effects of human neuronal aging might be linked to our long lifespan and are therefore not evident in animal models.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"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.