Imagine a future where healthcare is no longer reactive but predictive–where doctors can intervene early, tailor treatments, and monitor your health based on a real-time, personalized model of your body. That’s the ambitious vision driving the Human Phenotype Project (HPP), led by Professor Eran Segal, Chair of the Department of Computational Biology at MBZUAI.

Unlike other large-scale health studies such as the UK Biobank or All of Us, HPP stands out for its unprecedented depth and diversity. According to Segal, each participant is assessed across more than 30 different modalities–including genomics, the microbiome, sleep patterns, glucose levels, imaging, lifestyle, and clinical data. These assessments are not one-offs; they’re repeated every two years over a 25-year timeline, creating a living, evolving picture of human health.

DETAIL LIES IN DATA

So far, over 28,000 individuals have signed up, with 13,000+ completing initial data collection. Each participant contributes an average of 19 million structured data points, which are fed into generative AI models. These models don’t just describe past health–they predict future ones, turning raw data into actionable health insights. “It’s a shift from descriptive epidemiology to personalized forecasting,” says Segal.

BLOOD SUGAR MAY BE ONLT HALF THE STORY

Already, the project is challenging conventional diagnostic standards. Segal points to glucose monitoring as a prime example: roughly 40% of participants with normal fasting glucose actually exhibited prediabetic spikes after meals–information that would be missed without continuous monitoring. Similarly, markers like cholesterol and creatinine were shown to vary significantly based on age and ancestry, suggesting that current diagnostic norms may be outdated.

TOWARD A DIGITAL TWIN FOR HEALTHCARE

One of the most groundbreaking insights from HPP is its potential to redefine aging. Segal explains that biological age–measured through clinical and molecular data—is a stronger predictor of disease risk than chronological age. People with “older” biological profiles often show higher levels of glucose, visceral fat, and inflammation, even if they’re the same age as healthier peers. “Different body systems age at different rates,” says Segal. “This lets us target the right problem, in the right person, at the right time.”

With data at this scale and sensitivity, ethics and privacy are paramount. Every HPP participant gives informed digital consent, and can withdraw at any time. Personal identifiers are stripped before data is used, and external researchers must apply for access under strict terms. Even AI-driven simulations are built using tokenized timelines, and results are only shared in aggregate, preserving privacy without compromising insight.

The long-term goal? To create AI-powered digital twins for each participant–a living model that can simulate how different treatments or lifestyle changes might affect future health. “This isn’t just another data project,” says Segal. “It’s a blueprint for the future of personalized medicine, where care is predictive, preventative, and truly personalized.”