Microscopy deep learning predicts virus infections and reveals mechanics of lytic-infected cells

Abstract

Imaging across scales reveals disease mechanisms in organisms, tissues, and cells. Yet, particular infection phenotypes, such as virus-induced cell lysis, have remained difficult to study. Here, we developed imaging modalities and deep learning procedures to identify herpesvirus and adenovirus (AdV) infected cells without virusspecific stainings. Fluorescence microscopy of vital DNA-dyes and live-cell imaging revealed learnable virus-specific nuclear patterns transferable to related viruses of the same family. Deep learning predicted two major AdV infection outcomes, nonlytic (nonspreading) and lytic (spreading) infections, up to about 20 hr prior to cell lysis. Using these predictive algorithms, lytic and non-lytic nuclei had the same levels of green fluorescent protein (GFP)-tagged virion proteins but lytic nuclei enriched the virion proteins faster, and collapsed more extensively upon laser-rupture than non-lytic nuclei, revealing impaired mechanical properties of lytic nuclei. Our algorithms may be used to infer infection phenotypes of emerging viruses, enhance single cell biology, and facilitate differential diagnosis of non-lytic and lytic infections.

Virus Infection Variability by Single-Cell Profiling

“…This is likely due to cell intrinsic variability rather than genetic changes in the progeny. The recent development of machine-learning algorithms allows us to extract nuclear features from fluorescence microscopy and interrogate the molecular mechanisms of cell-to-cell variability in virus egress [ 69 ]. Intriguingly, convolutional neural networks revealed surprising differences between nonlytic and lytic infected cells.…”

Machine Learning and Artificial Intelligence for the Prediction of Host–Pathogen Interactions: A Viral Case

“… 60 Other examples include understanding structure and function relationships with the pathogens 40 or time-lapse analysis. 45 , 60 …”

“…Indeed, unsurprisingly for pathogen image classification tasks both shallower and deeper CNNs outperform conventional RF and MLP ML algorithms in metrics like F1 often by more than 30–40%. 47 , 60 …”

“…57 Downstream tasks (Figure 2B) for which ML is employed on such data typically include pathogen image segmentation 49 (often using a variety of the U-Net architecture 58 ), HPI events or virus classification from full or cropped field-of-view, 47,48,50 pathogen object detection 59 or infection detection. 60 Other examples include understanding structure and function relationships with the pathogens 40 or time-lapse analysis. 45,60 Being strongly related to the field of computer vision, HPI image analysis remains, thus far, strongly dominated by the CNN DL algorithms.…”

“…60 Other examples include understanding structure and function relationships with the pathogens 40 or time-lapse analysis. 45,60 Being strongly related to the field of computer vision, HPI image analysis remains, thus far, strongly dominated by the CNN DL algorithms. Indeed, unsurprisingly for pathogen image classification tasks both shallower and deeper CNNs outperform conventional RF and MLP ML algorithms in metrics like F1 often by more than 30-40%.…”