Today's Edition
MAYS IBRAHIM (ABU DHABI)
By tracking thousands of people over decades and analysing millions of biological signals in real time, scientists are collecting the kind of longitudinal data needed to move from reactive treatment to earlier risk detection.
Medical care has traditionally relied on measurements that offer only a brief snapshot of a person's health: a blood test, a clinic visit, or a one-time genetic scan.
But what if health could be tracked as a continuous story rather than a series of isolated images?
That is the ambition behind the Human Phenotype Project (HPP), a large-scale international study designed to follow individuals for up to 25 years to understand how health evolves and how disease begins long before symptoms appear.
In an interview with Aletihad, Eran Segal, Acting Dean of the School of Digital Public Health and Professor of Computational Biology at the Mohamed Bin Zayed University of Artificial Intelligence (MBZUAI), outlined how this project aims to move medicine from reaction to prediction.
Beyond DNA: Capturing Life in Motion
While genetics has dominated precision medicine headlines over the past two decades, Segal argues that DNA tells only part of the story.
"While a person's genome is the static blueprint of DNA they are born with, their phenotype consists of observable clinical, biological, and behavioural traits that emerge as genes interact with environment and lifestyle over time," he explained.
The HPP treats genetics as just one of more than 30 layers of data, according to Segal.
Researchers collect continuous glucose monitoring data, microbiome profiles, molecular markers, imaging, wearable sensor outputs, diet and sleep logs, and immune signals -creating one of the most deeply phenotyped datasets.
Unlike traditional biobanks that capture health at a single point in time, the HPP tracks participants continuously, which allows scientists to observe subtle biological changes that static measurements often miss.
"By capturing how health evolves day by day, the project enables a more complete and actionable form of precision medicine that can identify risk earlier and support prevention rather than late-stage treatment," Segal noted.
The project has so far enrolled approximately 28,000 participants, with more than 13,000 completing comprehensive assessments.
To address privacy concerns, Segal pointed out that all identifying details are removed before analysis, and advanced computational methods, including self-supervised learning, allow AI systems to detect patterns without relying on personally identifiable labels.
Early Detection and Tailored Prevention
One of the project's most significant findings so far indicates that standard screening tools miss a substantial proportion of at-risk individuals, according to Segal. He said that in one analysis, 40% of individuals with normal fasting glucose levels were reclassified as prediabetic when monitored continuously using glucose sensors.
Standard blood tests, researchers found, failed to detect substantial metabolic instability that only became visible through real-time monitoring.
The implications are significant, particularly in regions grappling with high rates of diabetes and heart disease.
Segal noted that cardiometabolic conditions often develop gradually, shaped by complex interactions between biology, behaviour, and environment.
By identifying risk earlier, researchers believe prevention could replace late-stage treatment as the dominant model of care.
Beyond early detection, the project is reshaping how diseases themselves are classified.
Rather than relying on broad labels, HPP researchers are identifying mechanistic subtypes based on underlying biological drivers. This approach could lead to more targeted therapies, personalised prevention plans, and tailored follow-up schedules, Segal noted.
Another surprising discovery by the HPP has been the outsized role of the microbiome, according to Segal. Researchers found that gut bacteria can influence weight loss and metabolism more strongly than genetic variation alone.
The microbiome was also shown to regulate hundreds of metabolites circulating in the bloodstream, some of which reach the brain and may influence neurological conditions, he added.
Powered by Artificial Intelligence
Artificial intelligence is central to the HPP's ability to process such vast and complex data collected over long periods of time, Segal noted.
The HPP integrates more than 30 phenotypic layers into AI systems capable of detecting patterns invisible to traditional statistical tools, he explained.
One such system, GluFormer – recently published in the journal Nature – analysed over 10 million glucose measurements from more than 10,000 participants. Using just two weeks of continuous glucose monitoring data, the AI model outperformed standard clinical markers such as HbA1c in predicting long-term cardiometabolic risk.
According to Segal, some insights from the HPP are already informing research settings.
Within one to three years, findings could begin shaping preventive strategies more broadly, he noted. Integration into everyday clinical practice, however, will likely take five to 10 years, pending validation across diverse populations.
