In a groundbreaking advancement that merges biological science with cutting-edge artificial intelligence, researchers from Edith Cowan University (ECU), in collaboration with Royal Prince Alfred Hospital in Sydney and Shantou University Medical College in China, have unveiled a pioneering method to measure biological age with unprecedented accuracy. Distinct from chronological age, which merely counts the years since birth, biological age offers a dynamic portrayal that assesses how well—or poorly—the body is aging based on molecular and cellular markers. This innovation promises to transform how we understand aging and age-related diseases, potentially ushering in a new era of personalized medicine and preventative healthcare.
At the heart of this breakthrough lies the integration of two complex biological data sets: the IgG N-glycome and the blood transcriptome. The IgG N-glycome pertains to the intricate sugar structures covalently attached to immunoglobulin G (IgG) antibodies. These glycan modifications play key roles in immune function and have been shown to evolve with age, reflecting immune system remodeling and systemic physiological changes. On the other hand, the transcriptome captures a snapshot of active gene expression within blood cells at any given moment, offering a dynamic overview of cellular activity and responses to internal and external stimuli. By examining these two layers together, the researchers aimed to encapsulate a more holistic fingerprint of biological aging.
Developing a tool capable of synergizing these complex data types was a formidable challenge, expertly addressed through the application of Deep Reinforcement Learning, a sophisticated form of artificial intelligence where algorithms iteratively learn optimal decision-making strategies from interacting with the data environment. This approach led to the creation of an ageing clock dubbed “gtAge,” a model that comprehensively interprets multi-omics inputs to predict biological age. Dr Xingang Li, a leading co-author and Postdoctoral Research Fellow at ECU, elucidated that gtAge can predict chronological age with a remarkable 85.3% accuracy, substantially surpassing previous models that relied solely on either glycomic or transcriptomic data.
The implications of this enhanced precision are profound. The model calculates what is known as the “delta age,” the discrepancy between predicted biological age and actual chronological age. This delta age correlates significantly with well-established markers of health and aging, including cholesterol levels, glucose metabolism, and other cardiovascular and metabolic indicators. Such linkage suggests that gtAge not only measures aging but also provides an actionable metric related to an individual’s real health risks, offering a potential early diagnostic tool for age-associated diseases.
Dr Li highlighted the critical limitation of relying exclusively on chronological age, pointing out its inability to capture the heterogeneity in aging observed across individuals. While some people experience pronounced physical and cognitive decline in their 60s or 70s, others maintain robust health well into nonagenarian years. This variation is attributable to differences in biological age driven by genetics, lifestyle, nutritional status, and disease history, emphasizing the need for a metric like gtAge that reflects these nuanced factors.
The development of gtAge also underscores a triumph of interdisciplinary collaboration. ECU’s Dr Syed Islam, a Senior Lecturer in Computer Science, led the AI methodology. His team engineered a custom AI tool named “AlphaSnake,” which harnesses Deep Reinforcement Learning to intelligently select the most informative features from the multi-omics datasets, avoiding the traditional pitfalls of naïvely merging heterogeneous data. The algorithm effectively navigates the complex biological landscape, balancing signal extraction while minimizing noise and redundancy, thus optimizing the age prediction model.
Testing the model rigorously, the researchers applied gtAge to a cohort of 302 middle-aged adults participating in the Busselton Healthy Ageing Study in Western Australia. This study population provided a valuable landscape to evaluate the tool’s robustness across a typical demographic range. Findings demonstrated that gtAge not only reflected chronological age with high fidelity but also linked with biological markers indicative of health status, reinforcing its potential clinical utility.
In context, Australia’s population dynamics—marked by rising elderly demographics—amplify the relevance of such a tool. The ability to assess biological age precisely allows healthcare practitioners to identify patients at elevated risk of age-dependent disorders earlier, enabling timely intervention strategies that could delay or prevent disease onset. Dr Islam emphasized the prospective public health benefits, where early lifestyle modifications informed by biological age measurements could markedly improve quality of life and reduce healthcare burdens.
Importantly, the concept of an aging clock is not new, yet previous iterations struggled with limited accuracy, often due to reliance on single data types or insufficient integration methods. The gtAge clock sets a new benchmark by leveraging multi-omics integration facilitated through a novel AI framework, thus providing a richer, more accurate picture of aging biology that captures the multifaceted nature of the process.
Beyond predicting age, this multifactorial approach opens the door for uncovering mechanisms that drive aging at a molecular level. The integration of glycomic and transcriptomic data provides insights into immune modulation, inflammatory status, and genetic regulation affecting aging pathways. Such mechanistic understanding could inform drug discovery, therapeutic targeting, and the design of personalized anti-aging interventions.
Looking forward, the research team envisions expanding the utility of gtAge through larger, more diverse population studies and longitudinal tracking to monitor how biological age changes over time in response to interventions. This could enrich its predictive power and verify its role as a dynamic health biomarker. Furthermore, integrating additional omics layers, such as proteomics or metabolomics, may refine and enhance the model’s sensitivity and specificity.
The study detailing this advance, titled “Deep Reinforcement Learning–Driven Multi-Omics Integration for Constructing gtAge: A Novel Aging Clock from IgG N-glycome and Blood Transcriptome,” was published in the journal Engineering on August 19, 2025. The authors’ transparent declaration asserts no competing financial interests, affirming the integrity of their findings.
In summary, this transformative work represents a milestone in aging research and precision medicine. As technologies converge and sophisticated AI models emerge, tools like gtAge provide an empowering lens for clinicians and individuals alike to understand biological aging beyond the passage of time. By translating complex biological data into meaningful health insights, this innovation holds the promise of fostering healthier lifespans and reshaping ageing from an inevitable decline to a manageable, informed journey.
Subject of Research: Cells
Article Title: Deep Reinforcement Learning–Driven Multi-Omics Integration for Constructing gtAge: A Novel Aging Clock from IgG N-glycome and Blood Transcriptome
News Publication Date: 19-Aug-2025
Web References:
https://www.sciencedirect.com/science/article/pii/S2095809925004837?via%3Dihub
http://dx.doi.org/10.1016/j.eng.2025.08.016
References:
Li, X., Islam, S., Xia, Y., Baten, A., Tan, X., & Wang, W. (2025). Deep Reinforcement Learning–Driven Multi-Omics Integration for Constructing gtAge: A Novel Aging Clock from IgG N-glycome and Blood Transcriptome. Engineering. https://doi.org/10.1016/j.eng.2025.08.016
Keywords:
Biological age, Aging clock, IgG N-glycome, Blood transcriptome, Deep reinforcement learning, Multi-omics integration, Artificial intelligence, Machine learning, Precision medicine, Age-related diseases
Tags: age-related disease understandingartificial intelligence in healthcarebiological age measurementbiological science breakthroughsblood transcriptome studiesEdith Cowan University researchIgG N-glycome analysisimmune system aginginnovative aging researchmolecular markers of agingpersonalized medicine advancementspreventative healthcare strategies
