Researchers have developed a multi-species computational framework to estimate biological age and predict mortality rates using gene expression profiles across diverse mammalian tissues.

研究人員開發了一個多物種計算框架，利用多種哺乳類組織的基因表達譜來估算生物年齡並預測死亡率。

The study utilized a comprehensive meta-dataset comprising over 11,000 transcriptomes from mice, rats, crab-eating macaques, and humans, spanning more than 25 distinct tissues. By employing elastic net and Bayesian ridge regression models, the investigators constructed 'relative' transcriptomic clocks. These instruments demonstrate a high degree of precision in predicting chronological age and, crucially, expected mortality, thereby surpassing the utility of standard chronological clocks which often fail to capture the effects of lifespan-extending interventions.

該研究利用一個全面的元數據集，包含來自小鼠、大鼠、食蟹獼猴和人類的 11,000 多個轉錄組，涵蓋 25 種以上不同的組織。研究人員透過採用彈性網路（elastic net）和貝葉斯脊回歸（Bayesian ridge regression）模型，構建了「相對」轉錄組時鐘。這些工具在預測實際年齡以及至關重要的預期死亡率方面表現出高度精準，因此超越了標準時鐘的效用，因為後者往往無法捕捉延長壽命干預措施的效果。

Analysis of the molecular data revealed a conserved architecture of ageing across species. Specifically, the upregulation of genes such as CDKN1A and LGALS3, alongside the downregulation of Col1a1 and Nrep, emerged as universal hallmarks of mammalian senescence. The researchers further decomposed these signatures into co-regulated functional modules, including interferon signaling, mitochondrial function, and chromatin modification. This modularity allows for the quantification of pathway-specific ageing; for instance, chronic pathologies primarily accelerate the inflammatory module, whereas caloric restriction modulates metabolic and mitochondrial components.

分子數據分析顯示，不同物種之間存在保守的衰老結構。具體而言，CDKN1A 和 LGALS3 等基因的上調，以及 Col1a1 和 Nrep 的下調，成為哺乳類衰老的普遍特徵。研究人員進一步將這些特徵分解為共調節的功能模組，包括干擾素信號、線粒體功能和染色質修飾。這種模組化使得量化特定路徑的衰老成為可能；例如，慢性病主要加速炎症模組，而熱量限制則調節代謝和線粒體組分。

Furthermore, the framework's validity was corroborated through the application of single-cell RNA sequencing and the analysis of rejuvenation models. The data indicate that pro-mortality expression shifts occur across nearly all cell types, including stem cell populations. The observed transcriptomic deceleration during early embryogenesis and cellular reprogramming suggests a potential for molecular rejuvenation. In human cohorts, these transcriptomic biomarkers demonstrated a predictive capacity for all-cause mortality comparable to second-generation epigenetic clocks, while providing superior mechanistic interpretability through functional gene annotation.

此外，透過應用單細胞 RNA 定序和分析回春模型，證實了該框架的有效性。數據表明，幾乎所有細胞類型（包括幹細胞群）都會發生促進死亡的表達轉移。在早期胚胎發育和細胞重編程過程中觀察到的轉錄組減速，暗示了分子回春的可能性。在人類隊列中，這些轉錄組生物標記對全因死亡率的預測能力與第二代表觀遺傳時鐘相當，同時透過功能基因註釋提供了更優越的機制解釋力。

The developed transcriptomic clocks provide a generalizable, multi-species tool for quantifying molecular age and mortality risk across various tissues and cell types.

所開發的轉錄組時鐘提供了一個可推廣的多物種工具，可用於量化各種組織和細胞類型的分子年齡與死亡風險。