A groundbreaking study published in Science Advances unveils a sophisticated method for predicting biological aging (BA) using deep neural networks (DNN). Unlike chronological aging, BA focuses on the intricate biological processes that govern the decline of bodily functions over time. The study leverages pathways of steroidogenesis, offering a more precise approach to understanding aging and its associated health risks. By identifying key biomarkers such as cortisol, this innovative model aims to redefine how we perceive and measure aging.
In recent years, researchers have sought to move beyond traditional phenotypic indicators like grip strength or lung capacity, which often lack standardization and precision. This new DNN model incorporates biochemical processes within critical steroid pathways, enhancing its interpretability and predictive power. The study utilized data from 100 healthy participants aged between 20 and 73, alongside a validation cohort of 50 individuals aged 40 to 59, ensuring robust results.
The foundation of this research lies in the complex biological process of aging, characterized by the accumulation of molecular and cellular damage leading to functional decline. While chronological aging is simply a measure of time passed, biological aging provides deeper insights into the physiological changes occurring within the body. Traditional methods for assessing BA have limitations, often relying on phenotypic indicators that fail to capture specific metabolic or physiological pathways contributing to aging.
To address these shortcomings, the study developed a DNN model based on pathways of steroidogenesis. Researchers quantified 22 steroids in 150 individuals using liquid chromatography-tandem mass spectrometry (LC-MS/MS), grouping them by sex and designation for training or independent validation. A custom-designed loss function was incorporated to account for the progressive heterogeneity of aging, significantly improving the model's accuracy.
Key findings revealed significant differences in steroid profiles between sexes, with corticosteroid and sex hormone pathways playing crucial roles in influencing aging trajectories. Cortisol emerged as a pivotal biomarker, demonstrating a positive correlation with biological aging due to its involvement in processes such as gluconeogenesis and inflammation. For women, steroids like 17-OH-P4, COR, COS, and TH-COL positively influenced BA, whereas men experienced effects tied to pregnenolone and testosterone levels.
Interestingly, only male smokers exhibited an accelerated aging trajectory compared to non-smokers, potentially linked to lower smoking frequency among women. These insights underscore the importance of considering environmental and behavioral factors when refining the model for broader applications.
This cutting-edge DNN model not only captures the increasing heterogeneity of aging but also sheds light on the dynamic biological processes influenced by steroidogenesis. Future refinements may incorporate total cholesterol as a reference, preserving predictive accuracy even in datasets with fewer steroid measurements. Expanding the model's training with diverse datasets will further enhance its ability to examine sex-specific metabolic pathways and their variations with age.
By integrating advanced AI techniques with comprehensive biochemical analysis, this study paves the way for a deeper understanding of the aging process. Its implications extend beyond mere prediction, offering potential avenues for early disease detection and personalized healthcare strategies tailored to individual aging profiles.