(721g) Machine Learning for Improved Interpretability of High Dimensional Single-Cell Data: Impact of Host Factors on the Immune Landscape of Influenza Infection

Influenza poses a persistent health burden worldwide. In order to design equitable vaccines and therapeutics that are effective across all demographics, it is essential to better understand how host factors such as genetic background and aging affect the immune response to influenza infection. Cytometry by time-of-flight (CyTOF) represents a promising technique to characterize the single-cell immune landscape. However, the biological interpretation of large, high-dimensional CyTOF data remains a challenge. To improve data handling and interpretability, we developed a new analytical approach iGATE (in-silico gating annotating training elucidating) based on probabilistic support vector machine. With the capability to rapidly and accurately “gate” tens of millions of cells in silico into user-defined types, iGATE enabled us to track 24 canonical immune cell types in mouse lung over the course of influenza infection with regards to their abundance, activation, cytokine production, susceptibility to influenza infection and permissiveness to influenza replication. Applying iGATE to study the effects of host genetic background, we show that the lower survival of C57BL/6 mice compared to BALB/c is associated with a more rapid accumulation of inflammatory cell types and decreased IL-10 expression, suggesting an overactive inflammatory response is detrimental to survival. Further, our data demonstrate that the most prominent effect of aging is a defective T-cell response, leading to reduced survival of aged mice compared to their young counterparts. Finally, iGATE reveals that the 24 canonical immune cell types exhibit differential influenza infection susceptibility and replication permissiveness in vivo, but neither property varies with host genotype or aging. Software is available at https://github.com/UmichWenLab/iGATE.