The Temporal Challenge of Genes in Lifespan

Although the average human lifespan has increased by more than five years in the past two decades, the fundamental genetic reasons why some individuals live much longer and others die earlier remain unclear. Previous studies primarily focused on scanning for overall lifespan-associated genes at a single time point, assuming that their effects are constant throughout life. This approach overlooked the subtle, age-dependent regulatory mechanisms. In particular, the hypothesis that longevity genes like FOXO3 have protective effects in youth and that mTOR signaling inhibition has effects in old age, operate at different times, has not yet been experimentally verified. Against this backdrop, the research team posed the challenge: "If the effects of genes change over time, then age-specific genomic analysis is needed to capture this." However, simultaneously handling large-scale, age-specific cohorts and high-resolution transcriptomic data poses limitations for existing statistical methods due to the volume and complexity of the data. Therefore, the researchers designed a new Bayesian framework that treats age as a continuous variable and allows for dynamic modeling of gene-environment interactions.

Age-Specific Gene-Signal Association Analysis and Key Findings

The research team extracted age, time of death, and transcriptomic profiles from a 500,000-person British Biobank and a 200,000-person U.S. healthcare database, and performed GWAS and transcriptomic association analyses every five years. The analysis revealed that the IGF-1 (insulin-like growth factor 1) signaling pathway and the inflammatory cytokine IL-6 exhibit an inverted effect: they promote lifespan extension in youth but exacerbate cardiovascular risk in middle age. In particular, variants of the SIRT1 (sirtuin 1) gene were found to strengthen mitochondrial function and reduce oxidative stress in their 30s, but accelerate cellular damage by inhibiting autophagy in their 70s, demonstrating age-dependent duality. These dynamic effects could not be predicted by existing fixed hazard models, and the new Bayesian age-varying model showed an average of more than 15% improvement in prediction accuracy. Furthermore, the researchers identified interesting differences in specific gene-environment interactions, such as the FOXO3 variant extending lifespan by more than 20 years in smokers but having little effect in non-smokers. Finally, they released an interactive map visualizing age-dependent gene effects, allowing researchers and clinicians to design personalized anti-aging strategies.

Implications and Future Prospects

The dynamic genetic effects demonstrated in this study challenge the conventional expectation of "predicting a lifetime with a single genomic test" and present a new paradigm that age-specific personalized genetic information interpretation is necessary. Pharmaceutical companies can now design age-segregated drug pipelines targeting FOXO3 and SIRT1, and Vein Bio and MediPlan are currently entering phase 2 clinical trials for anti-aging candidates targeting individuals in their 60s. Public health authorities have launched a pilot project to include gene-based personalized advice in national health screening programs using age-specific hazard models, which has the potential to expand the $2 trillion preventive healthcare market annually. In academia, the framework will be used to compare long-lived populations (e.g., Okinawa, Japan) and early-dying populations, in order to explore in greater depth how environment and genetics synergize. Furthermore, data scientists are accelerating efforts to develop AI-based predictive models using the publicly available age-varying genomic database and to integrate personalized lifespan extension strategies into digital health platforms. Ultimately, by understanding how genes dance with time, we will gain a new hope that "aging is not an inevitable fate, but a controllable process."