DrSpace: Disease-rating based on spatially resolved cancer micro-environment

1. Description

Aim to solve:

• Annotate tumor micro-environment:

  • Achieve distinct annotation of malignant cells and immune cells in the tumor microenvironment.
  • Enable precise annotation of immune cell subtypes.

• On the basis of annotating tumor micro-environment, we employ CellChat to construct a tumor micro-ecological network.

2. Installation

Installing Copykat from GitHub.

if (!requireNamespace("copykat", quietly = TRUE)) { 
    devtools::install_github("navinlabcode/copykat")
}

Installing CellChat from GitHub.

if (!requireNamespace("CellChat", quietly = TRUE)) { 
    devtools::install_github("jinworks/CellChat")
}

Installing DrSpace from GitHub.

if (!requireNamespace("DrSpace", quietly = TRUE)) { 
    devtools::install_github("QianChwnLyn/DrSpace")
}

3. Usage

Example data can be downloaded here. Make sure the data contains spatial images information and clusters columns.

library(DrSpace)
library(Seurat)
load("obj.rda")
Seurat::SpatialDimPlot(obj, pt.size = 1.5, label = TRUE, label.size = 3)

Predict disease data using Copykat.

copy_obj <- Copykat(obj = obj, cancer = "colon cancer", n_PC = 10, genome = "hg20")
Seurat::SpatialDimPlot(copy_obj[[1]], pt.size = 1.5, label = TRUE, label.size = 2, group.by = "type")

Perform cell type enrichment analysis and predict cell types on spatial transcriptomic data using SSEA.

num_list <- seq(100,1000,100)
pred_obj <- SSEA(obj_list = copy_obj, num_list, cancer = "colon cancer", population_size = 20000)
anno_obj <- pred_obj[[9]]
Seurat::SpatialDimPlot(anno_obj, pt.size = 1.5, label = TRUE,label.size =2, group.by = "predict_spot")
Seurat::SpatialDimPlot(anno_obj, pt.size = 1.5, label = TRUE,label.size =2, group.by = "predict_spot_sub")
Seurat::SpatialDimPlot(anno_obj, pt.size = 1.5, label = TRUE,label.size =2, group.by = "predict_cluster")

anno_obj_cancer <- subset(anno_obj,type == "colon cancer")
Seurat::SpatialDimPlot(anno_obj_cancer, pt.size = 1.5, label = TRUE, label.size = 2, group.by = "predict_spot")

Construct Spot-Spot Communication Network on spatial transcriptomic data using CellChat.

ssc_pre <- SSC(anno_obj,json_path = "../data/spatial/scalefactors_json.json")
pathways.show <- “IL6”
CellChat::netVisual_aggregate(cellchat, signaling = pathways.show, layout = "circle")
CellChat::netVisual_aggregate(cellchat, signaling = pathways.show, layout = "spatial", edge.width.max = 2, vertex.size.max = 1, alpha.image = 0.2, vertex.label.cex = 3.5)
CellChat::netAnalysis_signalingRole_network(cellchat, signaling = pathways.show, width = 8, height = 2.5, font.size = 10)

Reference

1. Gao, R Jin et al., Delineating copy number and clonal substructure in human tumors from single-cell transcriptomes. Nat Biotechnol. doi:10.1038/s41587-020-00795-2.

2. Suoqin Jin et al., CellChat for systematic analysis of cell-cell communication from single-cell and spatially resolved transcriptomics, bioRxiv 2023