How high-throughput microscopy can contribute to COVID-19 diagnosis

Biological and biomedical imaging technologies are essential for addressing many research questions related to Covid-19 and for diagnosis. A recent study conducted by the Heidelberg University Center for Integrative Infectious Disease Research, in collaboration with EMBL and the Heidelberg Collaboratory for Image Processing, reveals a procedure that combines high-throughput microscopy and semi-automated image analysis of virus-infected cells from a cell line to identify SARS-CoV-2 specific antibodies in human serum samples. Christian Tischer, scientist at EMBL’s Advanced Light Microscopy Facility (ALMF), part of Euro-BioImaging’s EMBL-Node, contributed to the image analysis procedure developed for this study. To learn more about how biological imaging technologies can contribute to COVID-19 diagnosis, read the preprint, “Microscopy-based assay for semi-quantitative detection of SARS-CoV-2 specific antibodies in human sera,” and don’t miss the image data which is openly available in the BioImage Archive. 

Contributing to COVID-19 diagnosis 

In this study, all assay development, sample preparation and microscopy were performed at Heidelberg University’s Infectious Diseases Imaging Platform (IDIP) in biosafety containment 2 and 3 conditions (BSL‑2 and 3). The EMBL-Node contributed to the development of an automated image analysis workflow as part of the semi-quantitative high-throughput microscopy-based assay. Its mission - in collaboration with the Kreshuk lab at EMBL - was image and data handling, analysis and visualisation. Specifically, the goal was to identify human serum that contained SARS-CoV-2 specific antibodies (one of the three major classes present in human serum, IgG, IgA and IgM), and to develop a computational approach to carry out the identification process for a high volume of samples. In other words, the basic approach was to use cells infected with SARS-CoV-2 as an analysis platform to test whether serum prepared from human blood samples contained antibodies reacting with the viral antigens present on the infected cells.

Data management 

The amount of data handled in the initial two procedures was large – up to 50.000 images on 50 plates. Data processing time was between 2-3 hours per plate. Over the course of the study, up to 4 TB of data were processed and analyzed. The ALMF’s extensive experience in the analysis of high-throughput microscopy screens, the image analysis expertise of the Kreshuk Lab, as well as EMBL’s high-performance computational resources were very important to this study. 

The assay was then deployed in a large-scale (5,000 volunteers) population immunity study in Baden- Württemberg, Germany, conducted by a consortium of regional hospitals. The objective was to test for virus immunity in pairs of children and parents before reopening local schools. High-throughput microscopy and semi-automatic image analysis were used to identify specific SARS-CoV-2 antibodies in the study patients. To find out more about the methods and results of this important study, read the preliminary study report.

An effective procedure for diagnostics 

The study showed that the sensitivity and specificity of COVID-19 diagnosis - as compared to an approved ELISA-based diagnostic test - is improved with high-throughput microscopy and an automated image analysis workflow. One of the advantages of the method is the presence of all viral antigens on the SARS-CoV2-infected cells, as compared to only one (or a selected few) in ELISA-based and similar diagnostic tests. As different persons may develop different spectra of antibodies against the virus, the presence of all viral antigens (instead of one or a few), reduces the chances for false negative readings. It provides a compelling argument for using microscopy for COVID-19 diagnosis – and for the serological analysis of other virus infections. 

Sharing COVID-19 data sets with other researchers

To ensure the data developed in this study benefits a maximum number of researchers, all raw microscopy images, intermediate segmentation and infected cell classification results from this study, as well as quality control and final score results are available in the BioImage Archive (http://www.ebi.ac.uk/bioimage-archive) under accession number S-BIAD24. We hope you will have a look!

To learn more and see full results, read the BioRxiv preprint: , “Microscopy-based assay for semi-quantitative detection of SARS-CoV-2 specific antibodies in human sera”

About Euro-BioImaging’s EMBL Node

The Euro-BioImaging EMBL-Node offers a collection of state-of-the-art microscopy equipment and image processing tools. This Node supports in-house scientists and visitors in using microscopy methods for their research and regularly organizes in-house and international courses to teach basic and advanced microscopy methods. The services provided include project planning, sample preparation, microscope selection and use, image processing and visualization. Through Euro-BioImaging, life scientists, regardless of their affiliation or area of expertise, can apply to access services from 21 internationally-renowned imaging facilities called Nodes.

About the BioImage Archive The BioImage Archive stores and distributes biological images that are useful to life-science researchers. Its development will provide data archiving services to the broader bioimaging database community. This includes added-value bioimaging data resources such as EMPIAR, Cell-IDR and Tissue-IDR.