TOrBCellSelection

Separate T and non-T cells and select T cells; or separate B and non-B cells and select B cells.

If all of your cells are T/B cells, do not set any configurations for this process.

In such a case, SeuratClusteringOfAllCells should not be used, and SeuratClustering will be clustering all of the cells, which are all T/B cells.

There are two ways to separate T and non-T cells; or B and non-B cells:

  1. Use the an expression indicator directly from the metadata.
  2. Use the expression values of indicator genes, and the clonotype percentage of the clusters.

You can also use indicator gene expression values only to select T/B cells by setting envs.ignore_vdj to true.

Input

  • srtobj: Seurat object file in RDS/qs2
  • immdata: Immune repertoire data file in RDS/qs2

Output

  • outfile: Default: {{in.srtobj | stem}}.qs.
    Seurat object file in qs2 format
  • outdir: Default: details.
    Output directory with details

Environment Variables

  • ignore_vdj (flag): Default: False.
    Ignore VDJ information for T/B cell selection.
    Use only the expression values of indicator genes if True.
    In this case, the Clonotype_Pct column does not exist in the metadata.
    If you want to use k-means to select T/B cells, you must have more than 1 indicator gene, and the first indicator gene in envs.indicator_genes must be a positive marker, which will be used to select the cluster with higher expression values as T/B cells.
  • selector: The expression passed to tidyseurat::mutate(is_TCell = ...) to indicate whether a cell is a T cell. For example, Clonotype_Pct > 0.25 to indicate cells with clonotype percentage > 25% are T cells.
    If indicator_genes is provided, the expression values can also be used in the expression. For example, Clonotype_Pct > 0.25 & CD3E > 0.
    If selector is not provided, a kmeans clustering will be performed on the expression values of indicator_genes and Clonotype_Pct, with K=2, and the cluster with higher clonotype percentage will be selected as T/B cells.

  • indicator_genes (list): Default: ['CD3E'].
    A list of indicator genes whose expression values and clonotype percentage will be used to determine T/B cells.
    The markers could be either positive, such as CD3E, CD3D, CD3G, or negative, such as CD19, CD14, CD68, for T cells. For B cells, markers such as CD19, MS4A1 (CD20), CD79A, CD79B could be used.

  • kmeans (type=json): Default: {'nstart': 25}.
    The parameters for kmeans clustering.
    Other arguments for stats::kmeans can be provided here. If there are dots in the argument names, replace them with -.

Examples

Use T cell indicator directly

If you have a metadata like this:

id Clonotype_Pct seurat_clusters
1 0.1 1
2 0.3 2
3 0.5 3

With the configuration below:

[TOrBCellSelection.envs]
selector = "Clonotype_Pct > 0.25"

The T cells will be selected as:

id Clonotype_Pct seurat_clusters is_TCell
1 0.1 1 FALSE
2 0.3 2 TRUE
3 0.5 3 TRUE

Use indicator genes

Let's say we set the indicator genes to ["CD3D", "CD3E", "CD3G"].

The mean expression values will be calculated for each cluster:

id Clonotype_Pct seurat_clusters CD3D CD3E CD3G
1 0.1 1 0.1 0.0 0.1
2 0.3 2 1.2 1.3 0.6
3 0.5 3 1.5 0.8 0.9

Then a kmeans clustering will be performed on the mean expression values of the indicator genes, together with Clonotype_Pct, with K=2.

id Clonotype_Pct seurat_clusters CD3D CD3E CD3G is_TCell
1 0.1 1 0.1 0.0 0.1 FALSE
2 0.3 2 1.2 1.3 0.6 TRUE
3 0.5 3 1.5 0.8 0.9 TRUE

kmeans

The cluster with higher clonoype percentage will be selected as T/B cells (is_selected = TRUE), and sent to SeuratClustering for further clustering and downstream analysis.