A. YAKIMOVICH¹, C. RACHMÜHL², A. RUDNICKA¹, D. SEILER¹, R. WITTE¹, V. ANDRIASYAN¹, Y. YAMAUCHI¹, A. S. ZINKERNAGEL², U. F. GREBER<sup>1

1</sup>Institute of Molecular Life Sciences, University of Zurich, Switzerland, ²Division of Infectious Diseases and Hospital Epidemiology, University Hospital Zurich, University of Zurich, Switzerland

Dozens of years of research in the area of microbiota have brought little understanding of how bacteria and viruses influence each other in the way they infect mammalian cells. This is in part due of the considerable complexity of viral and bacterial infections alone. Nonetheless, bacterial and viral co-infections severely impact on clinical disease development and outcome, and can have challenging therapeutic issues, especially since both viral and bacterial infections can be persistent or acute. To deconvolve the complexity of viral-bacterial co-infections, we have developed high content single-cell infection readouts. This can in principle be achieved by microscopy or fluorescence associated cell sorting analyses. While the former approach lacks scale, the latter lacks subcellular resolution and time-resolved information from live cells. Our novel method termed Microscopy-based Fluorescence pAthogen Co-infection Setting (µFACS) scores bacterial and viral infections of eukaryotic host cells at multi-parametric single- cell readouts in high-content. The µFACS assay employs bleeding edge machine learning (deep learning) approaches for the identification of specific infection phenotypes. It is augmented by open source image analysis and data mining software, which potentially allows other researchers to build upon our efforts. Using three major human respiratory viruses, adenovirus, influenza virus, rhinovirus, and Staphylococcus aureus bacteria we show how virus infection is influenced by persistent bacterial infection. This project will help enhance molecular understanding in the emerging field of viral and bacterial co-infections and immunity with medical relevance.