tomasdemeter / covid_scrna-seq

A molecular single-cell lung atlas of lethal COVID-19

This paper by Melms et al. presents a comprehensive analysis of the cellular landscape and interactions in the lungs of 19 individuals who died of COVID-19 and 7 control individuals. The authors used single-nucleus RNA sequencing (snRNA-seq) to profile about 116,000 nuclei from snap-frozen lung tissue samples collected within hours of death. They identified substantial alterations in cellular composition, transcriptional cell states, and cell-to-cell interactions in COVID-19 lungs, revealing the molecular mechanisms of tissue damage and impaired regeneration. They also inferred protein activity and ligand-receptor interactions to identify potential therapeutic targets for severe COVID-19.

The main findings of the paper are:

• The lungs from individuals with COVID-19 were highly inflamed, with dense infiltration of aberrantly activated monocyte-derived macrophages and alveolar macrophages, but had impaired T cell responses.
• Alveolar type 2 cells adopted an inflammation-associated transient progenitor cell state and failed to undergo full transition into alveolar type 1 cells, resulting in impaired lung regeneration.
• The authors identified expansion of pathological fibroblasts that contributed to rapidly ensuing pulmonary fibrosis in COVID-19.
• Monocyte/macrophage-derived interleukin-1β and epithelial cell-derived interleukin-6 were unique features of SARS-CoV-2 infection compared to other viral and bacterial causes of pneumonia.
• The authors also observed ectopic tuft-like cells in the lung parenchyma of COVID-19 patients, which may have a role in airway inflammation and tissue regeneration.

The paper provides a valuable resource for understanding the pathophysiology of lethal COVID-19 and developing novel therapeutic strategies.

scvi and scanpy

To reanalyse the data obtained from this paper, I have used mainly two tools: scvi and scanpy.

scvi is a probabilistic framework for single-cell omics analysis that leverages deep generative models. scvi can perform various tasks such as dimensionality reduction, batch correction, differential expression analysis, imputation, denoising, and integration of multi-modal data. scvi is implemented as a Python library that is compatible with scanpy.

scanpy is a scalable toolkit for analysing single-cell gene expression data. scanpy can perform preprocessing, visualization, clustering, trajectory inference, differential expression testing, gene set enrichment analysis, and simulation of gene regulatory networks. scanpy is also implemented as a Python library that integrates with scvi.

Using these tools, I have performed the following steps:

• I downloaded the raw count matrices from the Gene Expression Omnibus (GEO) database and loaded them into scanpy.
• I filtered out low-quality nuclei based on the number of genes, UMIs, mitochondrial reads, and doublets.
• I integrated the individual samples using scvi's scvi.data.setup_anndata and scvi.model.SCVI functions to create a latent representation that accounts for batch effects.
• I performed dimensionality reduction using scanpy's sc.tl.umap function to obtain a UMAP embedding of the integrated data.
• I performed clustering using scanpy's sc.tl.leiden function.
• I performed manual annotation of the resulting cell clusters based on their gene expression profiles.
• I performed differential expression analysis using scvi's scvi.model.SCVI.differential_expression function to compare gene expression between clusters or conditions.
• I performed gene set enrichment analysis using scanpy's sc.tl.rank_genes_groups function to identify enriched pathways or signatures in each cluster or condition.
• I performed Gene Ontology (GO) enrichment analysis and KEGG pathway analysis using the GO_Biological_Process_2023 and KEGG_2021_Human libraries.
• I performed ETV5 gene expression comparison between controls and COVID19 samples and determined statistical significance using Mann–Whitney U test.
• I performed gene signatures scoring using scanpy’s sc.tl.score_genes function. To calculate a score for each cell based on the average expression of a set of genes, normalized by the average expression of a set of control genes.
• I performed B cell heavy and light chain analysis with two different approaches. First approach is searching for top heavy and light coexpressed chains pairs. The second is searching for the heavy and light chains with the highest expression separately and then looking for the pairs.

: Melms JC et al. A molecular single-cell lung atlas of lethal COVID-19. Nature. 2021 Jul;595(7865):114-119. doi: 10.1038/s41586-021-03569-1. Epub 2021 Apr 29. PMID: 33915569; PMCID: PMC8084440.

: Lopez R et al. A joint model of unpaired data from scRNA-seq and spatial transcriptomics for imputing missing gene expression measurements. Bioinformatics. 2020 Aug 1;36(Suppl_1):i308-i317. doi: 10.1093/bioinformatics/btaa460. PMID: 32657414; PMCID: PMC7355239.

: Wolf FA et al. Scanpy: large-scale single-cell gene expression data analysis. Genome Biol. 2018 Apr 6;19(1):15. doi: 10.1186/s13059-017-1382-0. PMID: 29626926; PMCID: PMC5889549.

: Melms JC et al. A molecular single-cell lung atlas of lethal COVID-19. Gene Expression Omnibus (GEO). 2021 . Available from: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE157103