Comparison within Seurat object • Single Cell Pathway Analysis (SCPA)

As well as importing expression matrices into compare_pathways, you can also compare populations directly within a Seurat or SingleCellExperiment object in one function within SCPA. To do this, you just need the Seurat/SCE object that is annotated with the groups you want to compare.

Let’s load in a few packages.

library(SCPA)
library(tidyverse)
library(msigdbr)
library(Seurat)

Getting the dataset

We’re going to do the same comparison of Th1 and Tcm cells that we did in the quick start tutorial, again using all Hallmark gene sets. We’ll use the Szabo, P… Sims, P T cell dataset that you can download here.

t_cells <- readRDS("szabo_t_cell.rds")

Quick Look at the data

DimPlot(t_cells) +
  theme(aspect.ratio = 1)

Getting the gene sets

We need to define our gene sets, and we can do this through the msigdbr package that contains a bunch of biological gene sets that can be found here. Here we just take the Hallmark gene sets (using the “H” call to msigdbr), and use the format_pathways function to get them in a readable format for SCPA.

pathways <- msigdbr("Homo sapiens", "H") %>%
  format_pathways()

SCPA comparison

Now to compare the Th1 and Tcm populations, we can use the compare_seurat function within SCPA. Here, group1 defines the column name for your metadata, and group1_population defines two values within that column. N.B. We’re going to compare populations within a Seurat object here, but it’s the same set up for a SCE object using the compare_sce function.

scpa_out <- compare_seurat(t_cells,
                           group1 = "cell", 
                           group1_population = c("cd4 th1", "cd4 tcm"),
                           pathways = pathways)

# For faster analysis with parallel processing, use 'parallel = TRUE' and 'cores = x' arguments

And that’s it. Your results will be stored in your scpa_out object.

Comparing across more than 2 populations

If you wanted to compare more than 2 populations in a Seurat object e.g. multiple time points, you just need to supply more arguments to the group1_population call. For example, on a hypothetical dataset:

scpa_out <- compare_seurat(different_dataset,
                           group1 = "hour", 
                           group1_population = c("0", "12", "24"),
                           group2 = "cell",
                           group2_population = "t_cells"
                           pathways = pathways)

# For faster analysis with parallel processing, use 'parallel = TRUE' and 'cores = x' arguments