The discovAIR project ran from January 2020 to June 2022, including a 6-month no-cost extension to deal with the delays caused by the COVID-19 pandemic. Given the focus of discovAIR on the Human Lung Cell Atlas (HLCA), and the fact that the pandemic hit Europe shortly after the start of the project, discovAIR partners have addressed society’s need for more detailed insight into the cellular and molecular mechanisms underlying COVID-19. Early on in the pandemic, discovAIR partners coordinated an international effort within the Lung Biological Network of the HCA consortium to analyse data from healthy people to understand how SARS-CoV-2 (the virus that causes COVID-19) enters the cells of our body. These analyses identified the upper airways (nasal cavity) as the main site of entrance and transmission of the virus. This discovery was shared widely using scientific and public news outlets. It had immediate consequences for appropriate use of personal protection such as facemasks. In addition, discovAIR partners have contributed to several studies on the mechanisms of disease in patients with COVID-19, some of which have already been published. Finally, discovAIR has used SARS-CoV-2 to see how specific lung cells respond to viral infection.
We created new ways to analyse datasets generated in discovAIR. For instance, discovAIR has developed a pipeline for combining and visualising data from different studies. This pipeline was first piloted to look into how age, BMI or smoking affects our genes, making it easier for SARS-CoV-2 to enter the body. This helped to understand why some groups were more likely to develop COVID-19. Building on these efforts, discovAIR partners have developed a way of combining lung datasets using all genes rather than a small selection of genes. This expertise was critical to deliver the first draft of the HLCA. The successful launch of the HLCA was presented and assessed as one of the key HCA successes to date at the annual general meeting (June 2022, Vienna). It has put the HLCA as frontrunner of the various tissue atlases within the international HCA consortium.
To achieve full impact for the HLCA, it has been made available through several open platforms, such as FastGenomics (developed by discovAIR partner Commasoft), Cellxgene from the Chan Zuckerberg Initiative and the Single Cell Expression Atlas hosted at the European Bioinformatics Institute. The atlas can be freely explored or downloaded. Research groups in the community lacking the infrastructure to do this can use option on the FastGenomics platform to perform these analyses online. The discovAIR consortium has further enabled wide use of the HLCA as a reference for label transfer by Azimuth or CellTypist, two frequently-used open platforms for annotating cells according to published references. At this moment, the HLCA is considered the gold standard for cell type annotation in lung tissue datasets. The HLCA was presented at the American Thoracic Society meeting in May 2022 and will be presented at the European Respiratory Society meeting in September 2022, thereby reaching a large proportion of the respiratory community across the globe.
The discovAIR consortium has also created disease atlases for asthma, chronic obstructive pulmonary disease (COPD), pulmonary arterial hypertension, idiopathic pulmonary fibrosis, and COVID-19. These show the changes in how cells behave in different diseases. The discovAIR consortium has also generated the first perturbation atlas for lung diseases. This uses cells from healthy people or people with lung disease. It monitors how cells change when stimulated with proteins that signal to cells or with environmental triggers such as house-dust mite, SARS-CoV-2 or respiratory syncytial virus. Finally, we mapped the cell types and the changing cell states that we observed in the HLCA, the disease atlas, and the perturbation atlas onto a model of the lung. This allowed discovAIR to define local cellular neighbourhoods in healthy lung tissue, and the changes in these cells between healthy and diseased lungs.
