Skip to content

Commit

Permalink
updated Fig names and added new supplementary figure in response to r…
Browse files Browse the repository at this point in the history 
…eviewer 3's comment
  • Loading branch information
@WalterMuskovic
WalterMuskovic committed Nov 17, 2021
1 parent 62c6c1d commit 1be9d9c
Show file tree
Hide file tree
Showing 26 changed files with 376 additions and 270 deletions.
2 changes: 1 addition & 1 deletion README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@

Droplet-based single cell RNA-sequencing methods utilise microfluidics to encapsulate individual cells in water-in-oil droplets, dramatically increasing throughput compared to plate-based protocols. While encapsulating cells, droplets also capture cell-free "ambient" RNA, a complex mixture of transcripts released from damaged, stressed or dying cells. High concentrations of ambient RNA hinder identification of cell-containing droplets and degrade downstream biological interpretation, by violating the assumption that a droplet contains RNA from a single cell. The concentration of ambient RNA depends on the nature of the input cell suspension, being more abundant in freshly dissociated solid tissue and samples containing fragile cell types such as brain tissue.

We developed the [dropletQC R package](https://powellgenomicslab.github.io/DropletQC/ "DropletQC R package") which utilises a novel QC metric, the nuclear fraction, to identify droplets containing ambient RNA or damaged cells. The nuclear fraction quantifies the amount of RNA in a droplet that originated from unspliced pre-mRNA. Ambient RNA consists mostly of mature cytoplasmic mRNA and is relatively depleted of unspliced nuclear precursor mRNA. Droplets containing only ambient RNA will have a low nuclear fraction score while damaged cells, due to loss of cytoplasmic RNA, will tend to have a higher score than intact cells.
We developed the [DropletQC R package](https://powellgenomicslab.github.io/DropletQC/ "DropletQC R package") which utilises a novel QC metric, the nuclear fraction, to identify droplets containing ambient RNA or damaged cells. The nuclear fraction quantifies the amount of RNA in a droplet that originated from unspliced pre-mRNA. Ambient RNA consists mostly of mature cytoplasmic mRNA and is relatively depleted of unspliced nuclear precursor mRNA. Droplets containing only ambient RNA will have a low nuclear fraction score while damaged cells, due to loss of cytoplasmic RNA, will tend to have a higher score than intact cells.

The schematic below illustrates how the nuclear fraction score, in combination with total UMI counts, can be used to identify empty droplets and damaged cells.

Expand Down
154 changes: 67 additions & 87 deletions code/supplementary_figure_3.R
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,112 +2,92 @@

# title: Plot supplementary figures
# author: Walter Muskovic
# date: 2021-01-28
# description: In this script we are going to create the supplementary figures
# presented in the manuscript
# date: 2021-08-26
# description: In this script we are going to create an additional supplementary
# figure



# Import libraries -------------------------------------------------------------

suppressPackageStartupMessages({
library(Seurat)
library(tidyverse)
library(patchwork)
library(DropletQC)
#library(ggExtra)
#library(ggpubr)
library(ggpubr)
})



# Load data --------------------------------------------------------------------
# Load data ---------------------------------------------------------------

# Load expression data
load("data_track/qc_examples.Rdata")
pbmc <- Read10X("data/PBMC/outs/raw_feature_bc_matrix/")[,qc_examples$cell_barcode[qc_examples$sample=="PBMC"]]
gbm <- Read10X("data/GBM/outs/raw_feature_bc_matrix/")[,qc_examples$cell_barcode[qc_examples$sample=="GBM"]]
hl <- Read10X("data/HL/outs/raw_feature_bc_matrix/")[,qc_examples$cell_barcode[qc_examples$sample=="HL"]]
mb <- Read10X("data/MB/outs/raw_feature_bc_matrix/")[,qc_examples$cell_barcode[qc_examples$sample=="MB"]]

# Simplify unresolved clusters for plotting
qc_examples <- qc_examples %>%
# Get proportion of UMIs mapped to lncRNAs
lncRNA <- data.frame(sample=c(rep("GBM", sum(qc_examples$sample=="GBM")),
rep("HL", sum(qc_examples$sample=="HL")),
rep("PBMC", sum(qc_examples$sample=="PBMC")),
rep("MB", sum(qc_examples$sample=="MB"))),
cell_barcode = c(colnames(gbm), colnames(hl), colnames(pbmc), colnames(mb)),
MALAT1_NEAT1 = c(colSums(gbm[c("MALAT1","NEAT1"),]),
colSums(hl[c("MALAT1","NEAT1"),]),
colSums(pbmc[c("MALAT1","NEAT1"),]),
colSums(mb[c("Malat1","Neat1"),])))


# Add to qc_examples data frame
qc_examples <- merge(qc_examples, lncRNA, by = c("cell_barcode", "sample")) %>%
mutate(MALAT1_NEAT1_percentage = 100*(MALAT1_NEAT1/umi_count))



# Plot --------------------------------------------------------------------

# Change names for plotting
qc_examples <-
mutate(
cell_type = replace(
cell_type,
cell_type == "astrocyte_unresolved_1",
"astrocyte_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "astrocyte_unresolved_2",
"astrocyte_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_1",
"neuron_unresolved"
qc_examples,
flag = case_when(
flag == "damaged_cell" ~ "damaged cell",
flag == "empty_droplet" ~ "empty droplet",
TRUE ~ "cell"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_2",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_3",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_4",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_5",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_6",
"neuron_unresolved"
sample = case_when(
sample == "GBM" ~ "Glioblastoma",
sample == "HL" ~ "Hodgkin's Lymphoma",
sample == "PBMC" ~ "PBMCs",
sample == "MB" ~ "Mouse brain"
)
)

# Wrap cell type names for plotting
qc_examples <- mutate(qc_examples,
cell_type = str_replace_all(cell_type, "_", " ")) %>%
mutate(cell_type = str_wrap(cell_type, width = 15))



# Figure S3 --------------------------------------------------------------------

# Create plots showing how both the distributions of the nuclear fraction and
# UMI counts can be quite different for different cell types in the same sample

supp3 <- function(sample_name, plot_title){
p1 <- ggplot(filter(qc_examples, sample==sample_name),
aes(x=cell_type, y=nuclear_fraction_droplet_qc)) +
geom_violin()+
labs(title=plot_title,
x="Cell type",
y="Nuclear fraction") +
theme(axis.text.x = element_text(angle = 45, vjust = 0.5))

p2 <- ggplot(filter(qc_examples, sample==sample_name),
aes(x=cell_type, y=log10_umi_count)) +
geom_violin()+
labs(title=plot_title,
x="Cell type",
y="log10(UMI count)") +
theme(axis.text.x = element_text(angle = 45, vjust = 0.5))
return(p1 + p2)
}

# Save out plot
sample_names <- c("MB", "GBM", "PBMC", "HL")
plot_titles <- c("Mouse brain", "Glioblastoma", "PBMCs", "Hodgkin's lymphoma")
s1 <- map(1:4, function(x) supp3(sample_names[x], plot_titles[x]))
# Plot
p1 <- lapply(c("Mouse brain", "Hodgkin's Lymphoma", "Glioblastoma","PBMCs"), function(i){
filter(qc_examples, sample==i) %>%
ggplot(aes(x=flag, y=MALAT1_NEAT1_percentage, fill=flag)) +
geom_boxplot() +
facet_wrap(~sample) +
scale_fill_manual(values=c("damaged cell"="#88419d",
"empty droplet"="#2166ac",
"cell"="#b2182b")) +
theme(legend.position = "none") +
xlab("") +
ylab("MALAT1/NEAT1 % of UMIs")
}) %>%
ggarrange(plotlist = ., labels =letters[1:4], ncol = 2, nrow = 2)

# Save out figure
ggsave(filename = str_glue('figures/Figure_S3.png'),
plot = s1[[1]] / s1[[2]] / s1[[3]] / s1[[4]],
device = "png", width = 35, height = 30, units = "cm")
plot = p1,
device = "png",
width = 30, height = 20, units = "cm")

ggsave(filename = str_glue('figures/Figure_S3.pdf'),
plot = s1[[1]] / s1[[2]] / s1[[3]] / s1[[4]],
device = "pdf", width = 35, height = 30, units = "cm")
print("Finished creating supplementary figures")
plot = p1,
device = "pdf",
width = 30, height = 20, units = "cm")

113 changes: 113 additions & 0 deletions code/supplementary_figure_4.R
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,113 @@
# Script information -----------------------------------------------------------

# title: Plot supplementary figures
# author: Walter Muskovic
# date: 2021-01-28
# description: In this script we are going to create the supplementary figures
# presented in the manuscript



# Import libraries -------------------------------------------------------------

suppressPackageStartupMessages({
library(tidyverse)
library(patchwork)
library(DropletQC)
#library(ggExtra)
#library(ggpubr)
})



# Load data --------------------------------------------------------------------

load("data_track/qc_examples.Rdata")

# Simplify unresolved clusters for plotting
qc_examples <- qc_examples %>%
mutate(
cell_type = replace(
cell_type,
cell_type == "astrocyte_unresolved_1",
"astrocyte_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "astrocyte_unresolved_2",
"astrocyte_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_1",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_2",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_3",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_4",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_5",
"neuron_unresolved"
),
cell_type = replace(
cell_type,
cell_type == "neuron_unresolved_6",
"neuron_unresolved"
)
)

# Wrap cell type names for plotting
qc_examples <- mutate(qc_examples,
cell_type = str_replace_all(cell_type, "_", " ")) %>%
mutate(cell_type = str_wrap(cell_type, width = 15))



# Figure S3 --------------------------------------------------------------------

# Create plots showing how both the distributions of the nuclear fraction and
# UMI counts can be quite different for different cell types in the same sample

supp3 <- function(sample_name, plot_title){
p1 <- ggplot(filter(qc_examples, sample==sample_name),
aes(x=cell_type, y=nuclear_fraction_droplet_qc)) +
geom_violin()+
labs(title=plot_title,
x="Cell type",
y="Nuclear fraction") +
theme(axis.text.x = element_text(angle = 45, vjust = 0.5))

p2 <- ggplot(filter(qc_examples, sample==sample_name),
aes(x=cell_type, y=log10_umi_count)) +
geom_violin()+
labs(title=plot_title,
x="Cell type",
y="log10(UMI count)") +
theme(axis.text.x = element_text(angle = 45, vjust = 0.5))
return(p1 + p2)
}

# Save out plot
sample_names <- c("MB", "GBM", "PBMC", "HL")
plot_titles <- c("Mouse brain", "Glioblastoma", "PBMCs", "Hodgkin's lymphoma")
s1 <- map(1:4, function(x) supp3(sample_names[x], plot_titles[x]))
ggsave(filename = str_glue('figures/Figure_S4.png'),
plot = s1[[1]] / s1[[2]] / s1[[3]] / s1[[4]],
device = "png", width = 35, height = 30, units = "cm")
ggsave(filename = str_glue('figures/Figure_S4.pdf'),
plot = s1[[1]] / s1[[2]] / s1[[3]] / s1[[4]],
device = "pdf", width = 35, height = 30, units = "cm")
print("Finished creating supplementary figures")
11 changes: 4 additions & 7 deletions code/supplementary_figure_4.sh
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,11 +6,8 @@
#$ -q short.q
#$ -r yes
#$ -l mem_requested=4G
#$ -N supp_figure_2
#$ -N supp_figures


echo "Started running scripts to create supplementary figure 4 for the manuscript"

qsub get_cryo_microglia_data.R
qsub -hold_jid get_data align_cryo_microglia_data.R
Rscript plot_cryo_microglia_data.R
conda activate Seurat
echo "Started running script to create supplementary figure 4 for the manuscript"
Rscript supplementary_figure_4.R
76 changes: 0 additions & 76 deletions code/supplementary_figure_5.R
View file

This file was deleted.

Loading

0 comments on commit 1be9d9c

Please sign in to comment.