Resolving heterogeneity in Lymph Node Stromal Cells using high-dimensional analysis of non-optimized flow cytometry...

Area of Science:

• Immunology
• Cell Biology
• Bioinformatics

Background:

• Lymph Node Stromal Cells (LNSCs) are vital for maintaining lymph node structure and function, including immune response regulation.
• LNSCs hold potential for replicating lymph node functions *in vitro* due to their critical roles.
• The inherent heterogeneity of LNSCs presents significant challenges for their study and application.

Purpose of the Study:

• To demonstrate the difficulties in analyzing heterogeneous cell populations like LNSCs, particularly concerning changing population ratios and marker expression.
• To present a machine learning-based approach for more accurate and unbiased assessment of heterogeneous cell populations.
• To overcome limitations in flow cytometry analysis caused by experimental constraints and similar marker profiles.

Main Methods:

• Analysis of heterogeneous cell populations, focusing on Lymph Node Stromal Cells (LNSCs).
• Application of machine learning algorithms to identify and track cell population dynamics over time.
• Utilizing flow cytometry data while addressing challenges posed by similar marker expression and non-optimized controls.

Main Results:

• Demonstrated challenges in flow cytometry analysis of heterogeneous LNSCs due to population variability and marker overlap.
• Successfully applied machine learning to identify changing cell populations, reducing user bias in gating.
• Increased confidence in population identification by integrating multiple algorithms to overcome individual limitations.

Conclusions:

• Machine learning provides a robust method for analyzing complex, heterogeneous cell populations like LNSCs.
• This approach enhances the reliability of flow cytometry data analysis and reduces experimental bias.
• The findings support future research directions for understanding LNSC functions and *in vitro* applications.