Non-neural expression of SARS-CoV-2 entry genes in the olfactory epithelium suggests mechanisms underlying anosmia in COVID-19 patients

David Brann; Tatsuya Tsukahara; Caleb Weinreb; Darren W. Logan; Sandeep Robert Datta

Graph-based clustering was performed using the Louvain algorithm following the standard Seurat workflow. Cluster stability was analysed with Clustree on a range of resolution values (0.4 to 1.4), with 0.6 yielding the most stable set of clusters (72) . Overall, 26 clusters were identified, the smallest of which contained only 43 cells with gene expression patterns consistent with blood cell types, and was excluded from further visualisation plots. Clustering the two datasets separately yielded similar results. Moreover, the distribution of cells from each dataset across clusters was homogenous ( Figure S5 ) and the clusters corresponded previous cell class and subtype annotations (45) . As previously reported, a small cluster of excitatory neurons (cluster 13) contained neurons from the anterior olfactory nucleus. UMAP visualisations of expression level for cell class and cell type markers, and for genes coding for coronavirus entry proteins, depict log-normalized UMI counts. The heatmap in Figure 4B shows the mean expression level for each cell class, normalised to the maximum mean value. The percentage of cells per cell class expressing Ace2 was defined as the percentage of cells with at least one UMI.