Revolutionizing Healthcare Through Deep Phenotyping at MBZUAI

Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), the world's first dedicated AI research university in Abu Dhabi, United Arab Emirates, is at the forefront of transforming precision medicine. Central to this mission is the Human Phenotype Project (HPP), a groundbreaking longitudinal cohort study that deeply profiles thousands of participants to uncover hidden patterns in human biology. Launched under the leadership of Professor Eran Segal, Chair of Computational Biology at MBZUAI, the HPP integrates vast multimodal data with advanced AI to predict disease onset and progression long before symptoms appear.

This ambitious initiative reflects UAE's strategic push to lead in AI-driven healthcare innovation. By establishing a biobank and prospective cohort exceeding 10,000 participants—primarily Emiratis and residents—the project captures unprecedented depth in health data, setting new standards for personalized interventions in the region.

🔬 The Scope and Scale of the Human Phenotype Project

The HPP goes beyond traditional health studies by employing deep phenotyping, which involves comprehensive, repeated assessments of participants over time. Over 28,000 individuals have signed up, with more than 13,000 completing initial visits. Data collection encompasses:

• Medical history and family genetics
• Lifestyle, nutritional habits, and dietary logs
• Vital signs, anthropometrics (body measurements), and imaging like fundus and carotid ultrasound
• Blood tests for metabolome, transcriptome, and immune profiling
• Continuous glucose monitoring (CGM) generating millions of readings
• Sleep tracking via wearables
• Microbiome analysis from gut, oral, and vaginal samples

This multimodal approach allows AI models to simulate 'digital twins'—virtual replicas of individuals—for testing interventions. Early analyses have revealed novel microbiome signatures linked to breast cancer, inflammatory bowel disease (IBD), and endometriosis, as well as metabolic deviations in pancreatic cancer patients.

In the UAE context, where metabolic diseases like diabetes affect over 12% of the population, such insights are vital for tailored public health strategies.

GluFormer: A Transformer-Based AI Breakthrough in CGM Analysis

The crown jewel of HPP research is GluFormer, a generative foundation model published in Nature on January 15, 2026. Co-authored by MBZUAI's Eran Segal, Hagai Rossman, and President Eric Xing, alongside collaborators from Weizmann Institute and NVIDIA, GluFormer redefines continuous glucose monitoring (CGM).

Trained self-supervised on over 10 million CGM readings (every 15 minutes) from 10,812 non-diabetic HPP participants, the transformer architecture analyzes raw glucose dynamics rather than summaries like average levels. It forecasts:

• Diabetes onset up to 12 years ahead (66% of cases in top-risk quartile)
• Cardiovascular death risk (69% of events in top quartile, 0% in lowest)
• Clinical metrics like fasting glucose, visceral fat, liver attenuation, kidney function, and lipids up to 4 years

Validated across 19 diverse external cohorts, GluFormer surpasses the clinical gold standard, Glucose Management Indicator (GMI), and HbA1c tests.

How GluFormer Outperforms Traditional Metrics Step-by-Step

Conventional CGM relies on GMI, an average-based proxy for HbA1c, missing dynamic fluctuations. GluFormer processes time-series data through:

1. Input Encoding: Raw CGM traces tokenized into sequences capturing variability.
2. Self-Supervised Pretraining: Learns glucose patterns without labels, enabling generalization.
3. Fine-Tuning: Predicts future outcomes using HPP's 12-year AEGIS follow-up (580 adults).
4. Multimodal Extension: Integrates diet logs to simulate post-meal glucose responses.

"GluFormer’s ability to predict clinical outcomes up to four years in advance... provides unparalleled insight," says Eran Segal. In risk stratification, it identifies high-risk groups with precision unattainable by snapshots.

Beyond GluFormer: HPP's Broader Precision Medicine Insights

A July 2025 Nature Medicine paper from HPP highlighted phenotypic variations by age and ethnicity, urging personalized norms. Key findings include:

• CGM reclassifies prediabetes via day-to-day variability.
• Biological aging (via RNA-seq) predicts worse metrics like high triglycerides and HbA1c.
• Microbiome shifts in diseases; e.g., specific bacteria in breast cancer.
• Lifestyle impacts: Ultra-processed foods vs. Mediterranean diets alter trajectories.

AI frameworks like COMPRER predict cardiovascular risks from imaging. These pave the way for digital twins simulating diets or drugs.

MBZUAI's Trailblazing Researchers and Global Collaborations

Eran Segal, with dual appointments at MBZUAI and Weizmann, drives HPP's AI innovations. Eric Xing, MBZUAI President, emphasizes: "This is the dawn of true precision medicine." Hagai Rossman advances multimodal models.

Partnerships include Abu Dhabi Department of Health for disease prediction, Emirati Genome Program for genotypes, and NVIDIA for compute. Ties to UAE's healthcare ecosystem amplify impact.

For aspiring researchers, higher ed research jobs at MBZUAI offer cutting-edge opportunities in computational biology.

UAE's Strategic Role in Global AI Health Innovation

In the UAE, where diabetes prevalence nears 13%, HPP addresses national priorities. MBZUAI's integration with Emirati Genome enhances genotypic-phenotypic links, fostering proactive care. The project supports UAE's vision for AI leadership, with events like the 2026 HPP Conference partnering with Nature.

MBZUAI-DoH Abu Dhabi partnership accelerates AI models for public health.

Students and faculty contribute via PhD programs in computational biology, leveraging HPP data.

Challenges and Solutions in Scaling AI Phenotyping

Challenges include data privacy, ethnic diversity, and longitudinal retention. HPP solutions: ethical AI frameworks, UAE-centric cohorts expanding globally, and federated learning.

Benefits for stakeholders: Clinicians gain predictive tools; patients receive tailored plans; policymakers inform strategies. Real-world case: HPP's prediabetes reclassification could prevent thousands of UAE cases annually.

Future Outlook: Digital Twins and Beyond

HPP envisions personalized digital twins by 2030, simulating lifelong health under scenarios. Upcoming: Multimodal AI fusing all data for holistic predictions; microbiome therapeutics; global cohorts.

In UAE higher education, MBZUAI exemplifies AI's role in biology, inspiring UAE university jobs growth. Explore career advice for AI researchers.

Photo by Kelly Sikkema on Unsplash

Career Opportunities in UAE's AI Precision Medicine Boom

MBZUAI's HPP fuels demand for computational biologists, data scientists. PhD scholarships, internships like UGRIP attract global talent. Internal links: faculty positions, postdoc roles.

Check Rate My Professor for UAE AI faculty insights. UAE's AI ecosystem offers unparalleled growth.

For jobs, visit higher ed jobs, university jobs, or higher ed career advice. Post openings at /recruitment.