AI models predict biological age from small blood samples

Forget counting candles on your birthday cake. A new AI (AI) model developed by scientists at Osaka University in Japan can estimate biological age using only five drops of blood. By analyzing 22 key steroids, AI provides an individualized measure of how aging your body is, providing potential insights into health care and age-related diseases.

Aging is more than time passes. This is a complex process that is influenced by genetics, lifestyle and environment. Traditional methods for assessing biological age often rely on a wide range of biomarkers, such as DNA methylation and protein levels. However, these approaches can miss the complex hormonal networks that regulate our bodies. The Osaka University team focused on steroid hormones, compounds that are crucial in metabolism, immune function and stress responses.

“Our body relies on hormones to maintain homeostasis, why not use these as important indicators of aging?” says Dr. Qiuyi Wang, co-first author of the published study. Advances in science. This study introduces a deep neural network (DNN) model that incorporates steroid metabolic pathways, a new approach that enhances the biological interpretability of the model.

Instead of examining absolute steroid levels that can vary widely between individuals, the model examines the ratio between steroids. “Our approach reduces the noise caused by differences in individual steroid levels and allows models to focus on meaningful patterns,” explains co-first and corresponding author Dr. Zi Wang.

The team analyzed 22 steroids from 148 blood samples between 20 and 73 years old, with 98 samples for model training and 50 for verification. The AI ​​model captured the complex interactions between steroids and age, revealing that discrepancies between biological age and age tend to widen over time. Researchers liken this effect to the expansion of downstream rivers, showing how individual aging trajectories diverge over the years.

One of the most prominent findings is cortisol, a stress-related hormone. This study found that doubled cortisol levels result in approximately 1.5-fold increase in biological age. “Stress is often commonly discussed, but our findings provide concrete evidence that it has a measurable effect on biological aging,” says Professor Toshifumi Toshifumi Takao, a corresponding author and expert in analytical chemistry and mass spectrometry.

This model also revealed gender-specific variation in steroid metabolism. Clear patterns emerged between men and women, reflecting inherent biological differences. For example, estrogen-related steroids increased their influence in female models, whereas androgen-related steroids were more pronounced in male models. This underscores the importance of considering sex-specific hormone profiles in aging studies.

Interestingly, this study also investigated the effects of lifestyle factors such as smoking. The results showed that male smokers showed statistically significant acceleration in biological aging compared to non-smokers. This suggests that lifestyle choices can have measurable effects on the aging process, but this study acknowledges that more comprehensive data on other lifestyle factors is necessary for a more complete understanding.

Although this study offers promising insights, the authors point out some limitations. The relatively small sample size and lack of detailed lifestyle data may limit the generalizability of the findings. Furthermore, the model treats steroids in a static way and does not take into account circadian variations entirely. Future studies using larger cohorts and longitudinal data may help to further improve the model.

“This is just the beginning,” says Dr. Z. Wang. “We hope to expand our dataset and incorporate additional biological markers to further refine our models and unlock deeper insights into the mechanisms of aging.”

The potential applications for this AI-powered model are enormous. It could pave the way for more personalized health surveillance, early disease detection and customized wellness programs. The ability to assess “rate of aging” with simple blood tests could indicate an important development in preventive medicine.

Continuing advances in AI and biomedical research have accurately measured and even slowed down the increasingly feasible biological aging. For now, this study highlights the profound effect that hormones, especially stress-related hormones like cortisol, can have on our age.

A study entitled “Biological age prediction using a DNN model based on the pathway of steroid formation” Advances in science.