← Back to Home

Mapping Immunity in 80,000 People: UK Biobank Links Cell States to Long-Term Health

IMU Biosciences has begun analyzing blood from UK Biobank participants, linking immune cell data with genetic information, imaging and years of health records. The project aims to identify immune clues that precede disease, but further validation is needed before it can yield prediction tools for clinical use.

By SURL BioNews

A blood sample can reveal the current state of immune cells. Linking it to years of health records from the same person gives researchers an opportunity to ask: which changes reflect normal variation, and which might precede disease? IMU Biosciences announced on October 19 that it had begun immune profiling in September this year for an initial group of 80,000 UK Biobank participants, preparing to add more detailed immune information to this large research database.

According to IMU, the initial analysis includes 40,000 fresh blood samples and 40,000 cryopreserved samples. The project will link immune cell characteristics with existing genetic information, medical imaging, lifestyle data and long-term health records to explore how immune differences relate to disease risk, disease course and treatment response. The team hopes to eventually expand to all 500,000 participants, but the initial scale announced remains 80,000 people.

The focus of this work is to identify more cell populations simultaneously while making results comparable across large numbers of samples. IMU says its HD Immunology platform can resolve up to 2,500 immune cell subpopulations from a single blood sample, generating more than 100 million data points. These figures describe the depth of measurement; they do not mean that an equivalent number of disease markers has been identified. The research will combine Waters Corporation’s BD FACSDiscover cell analyzer and BD Rhapsody single-cell multiomics platform, with sequencing readouts provided by Ultima Genomics’ UG200 platform.

Funding will also support more detailed cell analysis. The UK’s Sovereign AI is providing £1.57 million through the Strategic Assets Programme. IMU further explained that this grant supports the first 10,000 samples and adds single-cell transcriptomic data to examine gene expression in individual cells. This allows researchers to investigate cells’ activity states as well as identify their types.

For artificial intelligence research, the practical application is to look for patterns between these cell states and subsequent health outcomes—for example, whether a particular immune profile is associated with developing disease later. Through standardized procedures, IMU hopes to reduce differences caused by laboratories or methods, preventing models from mistaking technical bias for biological signals. However, whether results from fresh and frozen samples are sufficiently comparable, and whether models can apply to other populations, remain questions that subsequent analyses must address.

What has been announced is a data-building project, rather than clinical results for a disease prediction tool. Ultima says the performance of UG200 used with BD Rhapsody was validated during the platform selection process. This type of technical validation does not demonstrate that a model can accurately predict disease. Even if associations between immune characteristics and disease are identified, causation cannot be inferred directly. Use in individual risk assessment or treatment selection would require confirmation through independent data and clinical studies.

Under the partners’ plans, the data generated will be returned to UK Biobank in stages and made available to approved researchers through existing procedures, without requiring a separate collaboration with IMU. This arrangement means the research value will depend not only on one company’s analytical capabilities, but also on whether different teams can test, reproduce and extend the findings. Whether an immune map of 80,000 people can lead to reliable health predictions will hinge on establishing links between cell measurements and real-world disease trajectories that can withstand validation.

References

  1. IMU Biosciences
  2. Ultima Genomics via PR Newswire
  3. Sovereign AI