hematopoiesis - case study

Reconstructing myeloid and erythroid differentiation using mouse scRNA data of Paul et al. (2015).

see also:

• https://scanpy-tutorials.readthedocs.io/en/latest/paga-paul15.html

• https://morris-lab.github.io/CellOracle.documentation/notebooks/03_scRNA-seq_data_preprocessing/scanpy_preprocessing_with_Paul_etal_2015_data.html

Mus musculus hematopoiesis single-cell data analysis example

Note

Since in the scRNA data only gene names are provided and not gene/transcript IDs, there is only partial overlap with the general transfer format (GTF) from mouse (Ensembl - release 110 - Mus_musculus.GRCm39.110.gtf.gz) and as such also with the OrthoFinder results (3044/3451). This is why it should be preferred to at least provide an additional column in the adata.var with a common GeneID from either NCBI or ensembl.

Please download the notebooks from here . Or please click below to view the content.

• Case study: re-analysis myeloid and erythroid differentiation of mouse (Mus musculus) single-cell data
  • Notebook file
  • Steps
  • Import libraries
  • Import oggmap python package submodules
  • Step 0 - run OrthoFinder to obtain orthogroups
  • Step 1 - get query species taxonomic lineage information
  • Step 2 - gene age class assignment (query species orthomap)
    • Gene age assignments per query species lineage node
    • Visualize number of species along query lineage and counts per gene age class
  • Step 3 - map OrthoFinder gene names and scRNA gene/transcript names
    • Import now, the scRNA dataset of the query species
    • Get an overview of observations
    • Helper functions to match gene names
    • Reduce isoforms to genes
  • Step 4 - get TEI values and add them to scRNA dataset
    • add TEI to adata object
  • Step 5 - downstream analysis
    • Boxplot gene age class per cluster
    • Use cluster names as cell-type
    • Boxplot gene age class per cell type
    • Plot relative expression per gene age class per cell type
    • Get partial TEI values to visualize gene age class contributions
    • Color UMAP/TSNE by TEI
      • Filtering
      • Normalization
    • Identification of highly variable genes
    • Log transformation
    • PCA and neighbor calculations
    • Cell clustering
    • Embedding the neighborhood graph
    • Color PAGA graph
      • UMAP