Metabolic Avatar
RESEARCH LINE
AI-Driven Precision Nutrition & Metabolic Modeling​
In a world where precision nutrition and personalized health are transforming how we prevent disease and optimize performance, traditional models of metabolism are no longer enough. The Metabolic Avatar represents a new frontier: an AI-driven system designed to model and predict your body’s unique metabolic responses in real time. What began as a Personalized Metabolic Avatar (PMA)—tailored to individual physiology, diet, and lifestyle—has evolved into a scalable Generalized Metabolic Avatar (GMA), capable of supporting entire populations with adaptive nutrition plans. Powered by data from wearable devices, dietary tracking, and cutting-edge machine learning algorithms, the Metabolic Avatar goes beyond calories. It integrates insights from gut microbiota, hormonal cycles, immune response, and even quantum-level biological dynamics to deliver a truly holistic view of human metabolism. This journey—from manual self-monitoring to intelligent metabolic forecasting—marks a paradigm shift in how we understand and manage health. Whether you are an athlete, a clinician, or someone seeking better well-being, the Metabolic Avatar is your gateway to a future of hyper-personalized, real-time metabolic optimization.
Next Step: Expanding QMA into the Generalized Quantum Metabolic Avatar (Q-GMA), combining quantum AI with large-scale metabolic predictions
Research Timeline:
The Evolution of the Metabolic Avatar
Milestones 2022–2025
The seven milestones of the Metabolic Avatar programme, developed at the Metabolic Intelligence Lab, Università Cattolica del Sacro Cuore in Rome.
2022: Genesis — The Foundation of Digital Self-Monitoring
The journey began with ArmOnIA, a web-based platform that integrated diet, physical activity, and body metrics into a single self-monitoring system. Unlike fragmented tracking tools, ArmOnIA offered real-time energy balance estimates, combining data from wearables and diet logs to give users a clear view of their metabolic status.
2023: Birth of the Personalized Metabolic Avatar (PMA)
The Personalized Metabolic Avatar (PMA) marked a major shift from static models to AI-driven metabolic forecasting. Using deep learning models (GRUs), PMA could learn from each user's unique metabolic data — diet, energy balance and activity — to predict personalized weight variations. With real-time input from wearable devices, the system dynamically adapted its predictions, paving the way for precision nutrition.
2023: Optimization & Real-World Scalability
After the success of PMA, the focus shifted to making the model faster, lighter and scalable for everyday use. A comparative study evaluated Deep Learning models (GRU, LSTM) against Statistical models (SARIMAX) for weight prediction. Results showed that deep learning consistently outperformed statistical methods, offering both higher accuracy and better computational efficiency, making PMA suitable for real-time deployment.
2024: Digital Biohacking & Sustainability
PMA evolved to embrace digital biohacking and environmental sustainability. Users began applying small, AI-guided dietary changes to improve metabolic efficiency and long-term health — this approach was defined as Digital Biohacking. In parallel, researchers extended PMA's capabilities to simulate the carbon footprint of different diets, enabling recommendations that were not only healthy but also environmentally responsible.
2024: Gut Microbiota & Precision Nutrition
The gut microbiota — a complex ecosystem of bacteria, fungi, viruses and archaea — emerged as a key regulator of metabolism, immune function and hormonal balance. This study demonstrated how microbiota-derived metabolites affect lipid and glucose metabolism, and influence the host's energy balance and inflammatory pathways. By integrating microbiome data into the PMA framework, the system became more adaptive and biologically precise.
2025: Generalized Metabolic Avatar (GMA)
The Generalized Metabolic Avatar (GMA) marks the transition from individualized metabolic models to scalable, population-wide predictions. Unlike PMA, GMA no longer requires personalized calibration — thanks to AI, it delivers instant, high-accuracy forecasts based on basic demographic data like age, gender and activity level. With processing time reduced from 12 s to 0.15 s, GMA enables real-time integration into large-scale health systems and apps.
2025: Quantum Metabolic Avatar (QMA)
The Quantum Metabolic Avatar (QMA) introduces a revolutionary leap in metabolic modeling through quantum computing. By integrating Quantum Echo State Networks (qESN), QMA achieves unprecedented accuracy and robustness, especially in noisy or data-scarce conditions. Quantum reservoir computing enables stable, high-fidelity predictions, reducing error margins by up to 76% compared to classical methods — paving the way for the next generation of AI-powered precision medicine.
Next Step: Expanding QMA into the Generalized Quantum Metabolic Avatar (Q-GMA), combining quantum AI with large-scale metabolic predictions
References
[1] Bianchetti, G.; Abeltino, A.; Serantoni, C.; Ardito, F.; Malta, D.; De Spirito, M.; Maulucci, G. Personalized Self-Monitoring of Energy Balance through Integration in a Web-Application of Dietary, Anthropometric, and Physical Activity Data. J. Pers. Med. 2022, 12, 568. https://doi.org/10.3390/jpm12040568
[2] Abeltino, A.; Bianchetti, G.; Serantoni, C.; Ardito, C.F.; Malta, D.; De Spirito, M.; Maulucci, G. Personalized Metabolic Avatar: A Data Driven Model of Metabolism for Weight Variation Forecasting and Diet Plan Evaluation. Nutrients 2022, 14, 3520. https://doi.org/10.3390/nu14173520
[3] Abeltino, A.; Bianchetti, G.; Serantoni, C.; Riente, A.; De Spirito, M.; Maulucci, G. Putting the Personalized Metabolic Avatar into Production: A Comparison between Deep-Learning and Statistical Models for Weight Prediction. Nutrients 2023, 15, 1199. https://doi.org/10.3390/nu15051199
[4] Abeltino, A.; Bianchetti, G.; Serantoni, C.; Riente, A.; De Spirito, M.; Maulucci, G. Digital Biohacking Approach to Dietary Interventions: A Comprehensive Strategy for Healthy and Sustainable Weight Loss. Nutrients 2024, 16, 2021. https://doi.org/10.3390/nu16132021
[5] Bianchetti, G.; De Maio, F.; Abeltino, A.; Serantoni, C.; Riente, A.; Santarelli, G.; Sanguinetti, M.; Delogu, G.; Martinoli, R.; Barbaresi, S.; et al. Unraveling the Gut Microbiome–Diet Connection: Exploring the Impact of Digital Precision and Personalized Nutrition on Microbiota Composition and Host Physiology. Nutrients 2023, 15, 3931. https://doi.org/10.3390/nu15183931
[6] A. Abeltino, C. Serantoni, A. Riente, M. De Giulio, S. Capezzone, R. Esposito, M. De Spirito, G. Maulucci. Transforming personalized weight forecasting: From the Personalized Metabolic Avatar to the Generalized Metabolic Avatar, Computers in Biology and Medicine, 2025. https://doi.org/10.1016/j.compbiomed.2025.109879.
[7] A. Abeltino, C. Serantoni, M. De Giulio, A. Riente, S. Capezzone, R. Esposito, M. De Spirito, G. Maulucci. Quantum Metabolic Avatar: a Digital Replica of Metabolism Enhanced by Quantum Algorithms, Expert Systems with Applications, 2026. https://doi.org/10.1016/j.eswa.2025.129045
[8] A. Riente, C. Serantoni, M. M. De Giulio, S. Capezzone, R. Esposito, M. De Spirito, G. Maulucci. Quantum Machine Learning in Diabetes and Metabolism: Applications and Translational Challenges, Computer Methods and Programs in Biomedicine, 2026. https://doi.org/10.1016/j.cmpb.2026.109609
Frequently asked questions
What is a Metabolic Avatar?
A digital replica of an individual metabolism, built from wearable and dietary data with deep-learning models and used to forecast weight and metabolic responses. It was developed at the Metabolic Intelligence Lab of Università Cattolica del Sacro Cuore in Rome.
What is the difference between PMA, GMA and QMA?
The Personalized Metabolic Avatar (PMA, 2023) is calibrated on each user, using gated recurrent units trained on their own diet, activity and energy balance. The Generalized Metabolic Avatar (GMA, 2025) removes the need for personal calibration and predicts from basic demographic data, cutting processing time from 12 seconds to 0.15. The Quantum Metabolic Avatar (QMA, 2025) runs on quantum echo state networks and reduces prediction error by up to 76 per cent in noisy or data-scarce conditions.
What data does the Metabolic Avatar need?
PMA needs continuous input from wearable devices plus a dietary log. GMA needs neither wearables nor tracking: age, gender and activity level are enough, which is what makes it deployable at population scale.
Is the Metabolic Avatar available to use?
The models are published and open to the research community through the papers listed on this page. The applied, everyday version of this work is Byo-me, the AI nutritional assistant the lab develops with Chirale S.R.L.
What is Digital Biohacking?
Small, AI-guided dietary changes applied progressively to improve metabolic efficiency and long-term health, guided by a Metabolic Avatar. The lab defined the term in a 2024 study in Nutrients that also simulated the carbon footprint of different diets.
The full publication list behind this work is on the Publications page.