كيف تتنبأ ساعة جينية عالمية بالشيخوخة وطول العمر

How transcriptomic clocks estimate biological age and predict mortality

Unlike chronological age, biological age reflects the actual functional state of cells and tissues. Transcriptomic clocks measure this biological age by quantifying the expression levels of hundreds to thousands of genes whose activity shifts in a predictable way over the life course .

The Tyshkovskiy and Gladyshev team developed their clocks by training machine learning models on a massive compendium of RNA‑seq data from rodents, primates, and humans, making them inherently multispecies and multitissue . The resulting models go beyond simply estimating chronological age. By integrating survival data for each sample, the researchers built a second generation of clocks trained directly on expected probability of death . These mortality‑trained clocks strongly predict how much biological aging has actually occurred and how soon death from any cause is likely to follow in mice, rats, macaques, and humans .

Crucially, the clocks capture both global aging signatures shared across tissues and tissue‑specific dysregulation. For example, an aging-related gene program involving inflammation and immune activation is consistently seen across brain, liver, kidney, and blood, whereas some metabolic shifts appear only in certain organs . This means a single blood sample can already give a meaningful readout of systemic biological age, but tissue‑specific clocks may be needed to track organ‑level aging .

Using the clocks to accelerate anti-aging intervention testing

A major bottleneck in geroscience has been the slow pace of testing whether an intervention truly extends healthy lifespan. Traditional longevity studies in mice can take three to four years and cost millions of dollars.

Transcriptomic clocks compress this timeline by providing a quick molecular readout of biological age. Instead of waiting for animals to die, researchers can take a tissue sample, measure the clock, and see whether the intervention shifted the transcriptome toward a younger state . Because the clocks are conserved across species, a drug that "rejuvenates" the transcriptome in mice can be directly benchmarked against human aging signatures in biobank data such as the UK Biobank, speeding the path from mouse studies to human relevance .

The study already showed that known lifespan‑extending interventions—including dietary restriction, rapamycin, and genetic models of longevity—are captured by the clocks, causing a younger transcriptomic profile . Researchers can now use the clocks as a screening tool to prioritize the most promising compounds for full lifespan studies, dramatically accelerating the search for effective anti-aging therapies .

Which gene pathways ramp up with age

Across all four species and the majority of tissues, aging consistently increased the activity of a core set of biological processes and specific genes. This conserved "up‑with‑aging" signature provides both a mechanistic map of what goes wrong in old cells and a set of biomarker genes for the clocks .

Inflammation and immune activation were the most universally elevated signals:

• Pathways related to interferon signaling, innate immunity, and the adaptive immune response were enriched in the age‑upregulated gene set .
• Whole‑blood transcriptome analyses independently validated that inflammatory and senescence‑related pathways rise with age and track with mortality risk .

Apoptosis and cellular senescence programs were also consistently up‑regulated across tissues, reflecting the accumulation of damaged and dying cells . Specific genes linked to senescence and cell‑cycle arrest, including CDKN1A and LGALS3, were not only increased with age in the transcriptomic data but also showed higher protein levels associated with mortality and multimorbidity in the UK Biobank .

Other hallmark up‑regulated processes included ribosomal and protein synthesis genes—likely a compensatory response to cellular stress—and certain immunoregulatory genes such as B2m .

In rodent‑focused transcriptomic atlases, the gene Gpnmb (glycoprotein non‑metastatic melanoma protein B) was identified as one of the most frequently up‑regulated age‑related genes across multiple tissues in both mice and rats . This gene is involved in lysosomal function and microglial activation and has now been repeatedly linked to brain and systemic aging.

Which gene pathways decline with age

The flip side of aging is a progressive loss of cellular maintenance and energy production. The transcriptomic clocks consistently picked up a down‑regulation signature centered on mitochondria and metabolism .

Mitochondrial function and oxidative phosphorylation were the most robustly declining pathways:

• Genes encoding components of the electron transport chain and mitochondrial ribosomes were down‑regulated in virtually every tissue and species examined .
• This loss of mitochondrial gene expression is considered a primary driver of age‑related energy failure and oxidative damage .

Broad metabolic processes also declined:

• Pathways related to oxidative phosphorylation, cellular respiration, and general energy metabolism were consistently suppressed with age .
• In parallel, processes such as wound healing, cell differentiation, and extracellular matrix synthesis were found to drop off as chronological age increases, reflecting a loss of tissue maintenance capacity .

At the single‑gene level, Asxl3 (Asxl transcriptional regulator 3) was repeatedly identified as one of the most frequently down‑regulated age‑associated genes across tissues in rodents . While the function of Asxl3 in aging is less well characterized than some other hits, its consistent decline makes it a useful clock component and a potential target for future functional studies.

Together, the transcriptomic hallmarks paint a clear picture: aging is not a random breakdown but a coordinated, evolutionarily conserved shift toward inflammation and away from energy production. The new multispecies clocks capture this shift in a quantitative, reproducible way, giving researchers a powerful tool to measure biological age, predict mortality, and test whether tomorrow's interventions can genuinely turn back the molecular clock .

The study discussed in this article, “Universal transcriptomic hallmarks of mammalian ageing and mortality,” was published in Nature on 27 May 2026 by Alexander Tyshkovskiy, Vadim N. Gladyshev, and an international team from Harvard Medical School and Brigham and Women’s Hospital, among others. The transcriptomic clock models featured in the paper are publicly available on Zenodo . A web‑based tool, TACO, hosted by the Gladyshev Lab, also allows interactive exploration of the dataset .