vis_LEC_tumor_skin

Visualize expression across different conditions (LECs in tumor and skin)

The experiment consist of 3 main factors of interest:

• sample origin: tumor, skin, lymph node

• main cell type class: leucocytes, LECs

• tumor type: YUMM, YUMMER

As a first check we are interested in the expression of specific marker genes/known marker and how they vary across combinations of these factors.

Preamble

library(dplyr)
library(tidyr)
library(ggplot2)
library(pheatmap)
library(patchwork)
library(Seurat)
library(stringr)
library(ggthemes)
#library(SingleCellExperiment)
library(gridExtra)

library(ggrepel)

Data objects

From the FGCZ we have different data objects:

1. Tumor LECs integrated (only tumor LECs, but have been integrated with skin samples) (data/scData_LEC_tumor.rds)

2. Tumor + skin LECs integrated (data/scData_LEC_tumor_skin.rds)

3. All skin + tumor cells

4. All lymphnode cells

Here we work with object 2 - Integrated tumor and skin LECs.

#### ----- change to local path to scData.rds object here ----- #####
# e.g. seurat <- readRDS("/home/project/data_folder/scData.rds)")
seurat<- readRDS(file.path("/Users/thomarin/Documents/PhD/Tumor project/Sequencing experiment/August 2023 first analysis shallow sequencing/rds file/scData_tumor_CD45.rds"))
#### --------------------------------------------------------- #####

# correct condition assignment!!
seurat$cond <- seurat[[]] |> 
  mutate(
    cond = case_when(
       str_detect(Sample, "YUMM[0-9]") ~ "YUMM",
       str_detect(Sample, "YUMMER") ~ "YUMMER",
       str_detect(Sample, "Skin") ~ "skin"
    )
  ) |> select(cond)

# check assignment
table(seurat$Sample, seurat$cond)

                      
                       skin YUMM YUMMER
  SkinLECs_Leukocytes1 1510    0      0
  SkinLECs_Leukocytes2 2351    0      0
  TumorYUMM1_1A           0  664      0
  TumorYUMM1_1B           0  586      0
  TumorYUMM2_1A           0 1886      0
  TumorYUMM2_1B           0 1652      0
  TumorYUMM5_2A           0 1829      0
  TumorYUMM5_2B           0 1603      0
  TumorYUMM6_2A           0 2380      0
  TumorYUMM6_2B           0 2267      0
  TumorYUMMER3_1A         0    0   1031
  TumorYUMMER3_1B         0    0    887
  TumorYUMMER4_1A         0    0   1442
  TumorYUMMER4_1B         0    0   1234
  TumorYUMMER7_2A         0    0   2015
  TumorYUMMER7_2B         0    0   1898
  TumorYUMMER8_2A         0    0   2593
  TumorYUMMER8_2B         0    0   2379

table(seurat$cond)

  skin   YUMM YUMMER 
  3861  12867  13479 

DefaultAssay(seurat) <- "SCT"

cluster_marker <- c("Cd3e", "Cd8a", "Cd4" ,"Tcf7", "Lef1", "Pdcd1", "Sell", "Tigit", "Cxcr4", "Ccr7", "Foxp3", "Ctla4", "Trdc", "Il12b", "Itgae", "Xcr1", "Cd207", "Notch2", "Itgax", "Itgam", "Il1a", "Cd80", "Cd19", "Prox1", "Pecam1", "Csf2rb", "Csf3r", "Il1r2")

cond_marker <- c("Cd3e", "Cd8a", "Cd4")

Overview

Distribution of conditions per cluster

cond_dat <- seurat[[]] |> group_by(cond, ident) |>  summarise(n_cells = n())

`summarise()` has grouped output by 'cond'. You can override using the
`.groups` argument.

ggplot(cond_dat, aes(fill=cond, y=n_cells, x=ident)) + 
  geom_bar(position="stack", stat="identity") + 
  scale_fill_tableau() + 
  theme_bw()

ggplot(cond_dat, aes(fill=cond, y=n_cells, x=ident)) + 
  geom_bar(position="fill", stat="identity") + 
  scale_fill_tableau() + 
  theme_bw()

Distribution of cells within umap

DimPlot(seurat, group.by = "ident", split.by = "cond", reduction = "umap") + ggtitle("")

#Almut please put number of clusters inside the plot

Cluster marker gene expression

Cluster	Genes	Cluster	Genes	Cluster	Genes
0		9		18
1		10		19
2		11		20
3		12		21
4		13		22
5		14		23
6		15		24
7		16
8		17

Dimplots

FeaturePlot(object = seurat, features = cluster_marker, ncol = 3)

Violinplots

VlnPlot(object = seurat, features = cluster_marker, ncol = 3, assay = "SCT", add.noise = F)

Dotplots

DotPlot(object = seurat, features = cluster_marker) + 
  theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust=1))

Conditional marker expression

Expression by condition as Dimplot

FeaturePlot(object = seurat, features = cond_marker, ncol = 3, split.by = "cond")

Expression as violin plot grouped by cluster

VlnPlot(object = seurat, 
        features = cond_marker, 
        assay = "SCT", 
        add.noise = F, 
        split.by = "cond",
        ncol = 1) + 
  plot_layout(guides = 'collect') &
  scale_fill_tableau() &
  theme_bw()

Expression as violin plot grouped by cluster

VlnPlot(object = seurat, 
        features = cond_marker,
        group.by = "cond",
        assay = "SCT", 
        add.noise = F,
        ncol = 2) + 
  plot_layout(guides = 'collect') &
  scale_fill_tableau() &
  theme_bw()