Tu185 - CyNET- a Network Analysis Framework for Holistic System Level Analysis of Immune Response

Joo Guan Yeo – Translational Immunology Institute (TII), SingHealth-DukeNUS; Sharifah Nur Hazirah – Translational Immunology Institute (TII), SingHealth-DukeNUS; Su Li Poh – Translational Immunology Institute (TII), SingHealth-DukeNUS; Nursyuhadah Binte Sutamam – Translational Immunology Institute (TII), SingHealth-DukeNUS; Martin Washer – Translational Immunology Institute (TII), SingHealth-DukeNUS; Gladys Ang – Translational Immunology Institute (TII), SingHealth-DukeNUS; Vasuki Ranjani – Translational Immunology Institute (TII), SingHealth-DukeNUS; Rachel Cheong – Translational Immunology Institute (TII), SingHealth-DukeNUS; Kim Huat Nicholas Khoo – Translational Immunology Institute (TII), SingHealth-DukeNUS; Jing Yao Leong – Translational Immunology Institute (TII), SingHealth-DukeNUS; Thaschawee Arkachaisri – Duke-NUS Medical School and Rheumatology and Immunology Service, KK Women's and Children's Hospital, Singapore; Salvatore Albani – Translational Immunology Institute (TII), SingHealth-DukeNUS

Abstract Text: Modern flow-cytometer and mass-cytometer (CyTOF) allow deep immune phenotyping and identification of various immune cell subsets in an unbiased manner. The established statistical procedures are used to find the perturbed subsets in comparative pathophysiological conditions. However, approaches that consider immune response at system levels are lacking. In real in-vivo conditions, various immune cell subsets interact and influence each other's functions. The coordinated interactions between these cell subsets determine whether the outcome is a normal physiological state or pathological condition. Established statistical procedures completely ignore the interactions between subsets and rely on statistically significant changes in the frequency of cell subsets. We developed CyNET (Cytometry Network)— an analysis platform based on a network science approach to understand the immune system holistically. In CyNET we quantified the systems level property and subsets level property. We used the CyNET to analyze the immune development with age. Peripheral blood cells from Healthy newborns (cord blood), adults (20 to 55 years), and old (70 years and above) human subjects were further used for validation using single-cell transcriptomics. We found that network edge density, degree centralization, and assortativity score reflect the maturation and development of the immune system at the age axis. Change in the centrality of the nodes (Immune subsets) better reflects the biological function than changes in frequency. In conclusion, our network analysis approach enables the summarization of immune response holistically and allows the identification of key or hubs of the immune cell network