Secreted protein signaling activity for distinct cell states from scRNA-Seq data • SecAct

This tutorial demonstrates how to infer secreted protein signaling activity in various cell states. For this demonstration, we employ an ovarian cancer scRNA-Seq study. Using the SecAct framework, we will estimate the signaling activity of secreted proteins for different CD4+ T cell states.

Load scRNA-Seq data

The scRNA-Seq data of CD4+ T cells is stored in a Seurat object. Three cell states are annotated in this dataset, i.e., naive, central memory, and helper 1-like.

library(SecAct)
library(Seurat)

# load scRNA-Seq data
dataPath <- file.path(system.file(package="SecAct"), "extdata/")
Seurat_obj <- readRDS(paste0(dataPath,"OV_scRNAseq_CD4.rds"))

# cell count
table(Seurat_obj@meta.data[,"Annotation"])

## CD4_central_memory          CD4_naive       CD4_Th1_like 
##                402                 28                358 

# show UMAP plot
Seurat::DimPlot(Seurat_obj, reduction = "umap", group.by = "Annotation")

Infer secreted protein activity

After loading scRNA-Seq data data, user can input Seurat_obj into SecAct.activity.inference to infer the signaling activities of secreted proteins for various cell states. Also, you need to claim which column in the meta matrix of Seurat object has cell state information. In this case, the column name is “Annotation” and thus assign cellType_meta = "Annotation". User can find the SecAct results in Seurat_obj @misc $SecAct_output $SecretedProteinActivity.

# infer activity; ~2 mins
Seurat_obj <- SecAct.activity.inference.scRNAseq(Seurat_obj, cellType_meta="Annotation")

# SecretedProteinActivity$zscore stores activity
act <- Seurat_obj @misc $SecAct_output $SecretedProteinActivity$zscore

# show activity
head(act)

The inferred protein activity is a relative measure. A positive value indicates high activity in the corresponding cell states, while a negative value signifies low activity.

Visualize activity score

User can visualize any secreted proteins of interest using SecAct.heatmap.plot. Here, we select the top active secreted proteins (SPs) for each cell state.

# select top secreted proteins
n <- 10
SPs.1 <- names(sort(act[,1],decreasing=T))[1:n]
SPs.2 <- names(sort(act[,2],decreasing=T))[1:n]
SPs.3 <- names(sort(act[,3],decreasing=T))[1:n]

SPs <- c(SPs.1, SPs.2, SPs.3)
# user can assign any secreted proteins to SPs

# subset act matrix
fg.mat <- act[SPs,]

# visualize in heatmap
SecAct.heatmap.plot(fg.mat)