Visualization

The vis workflow can take a MuData object from any of the other Panpipes workflows and visualize the following (where applicable):

  • for custom markers:

    • heatmaps

    • dotplots

    • embeddings such as PCA and UMAP coloured by the marker expression

  • for categorical variables:

    • barplots, e.g the number of cells per sample ID

    • stacked barplots, e.g. the number of cells in each cluster split by sample ID

    • embeddings such as PCA and UMAP coloured by the categorical variable

  • for continuous variables:

    • violin plots split by categorical variables, e.g doublet score per disease group

    • embeddings such as PCA and UMAP coloured by the continuous variable

  • for paired markers or metrics:

    • scatter plots, e.g. scatter plot of total_counts on the x axis and n_genes_by_counts on the y axis

For plots with custom markers follow the guidelines in Gene list formats to create the input gene lists.

Steps to run

  1. Activate conda environment conda activate pipeline_env

  2. Generate the config and log files with panpipes vis config and edit the pipeline.yml file

  3. (Optional) Prepare gene list files, more details here

  4. Run complete workflow with panpipes vis make full --local

The Visualization tutorial guides you through the visualization step by step.

Expected structure of MuData object

The vis workflow can take the MuData outputs from any of the other Panpipes workflows, but make sure not to ask the pipeline to produce visualisations using data that has not been computed, e.g. plotting UMAPs using a MuData object which does not contain anything in the X_umap slot.