Tu208 - Characterising Drivers of Human Immune Variation Using High Content High Throughput Cytometry and Machine Learning

Abstract Text: Immune system variation underpins diverse responses to therapies, vaccinations, and disease, yet remains poorly characterised across healthy populations. To address this gap, we conducted a comprehensive immune profiling study of 1,009 predominantly healthy individuals, integrating high-resolution spectral flow cytometry data with clinical blood biochemistry and extensive health and lifestyle questionnaires. Our novel machine learning (ML) platform provides real-time immune cell classification from spectral flow cytometry data, facilitating standardised, robust and reproducible characterization of over 2000 immune cell subsets.
Key findings from the study include age- and sex-associated immune signatures, revealing novel insights into immune aging and sexually dimorphic immune profiles. Further, by correlating immune phenotypes with clinical parameters, we identified hundreds of immunome-wide associations linking immune variation to health status and lifestyle factors.
The platform's scalability, enabled by industrialised sample processing and ML-driven analytics, addresses a critical need for population-scale immune profiling. The ability to normalise and interpret data from cryopreserved samples ensures compatibility with large biobanks. This capability sets the foundation for integrating immune data into broader clinical and genetic datasets to unlock new perspectives on human health and disease.
This study underscores the transformative potential of next-generation flow cytometry combined with ML analytics to generate high resolution immune profiling data at population-scale to enable precision immune medicine, patient stratification, and biomarker discovery.