-
Notifications
You must be signed in to change notification settings - Fork 10
/
jose_deg_and_enrichment_figures.R
241 lines (172 loc) · 9.44 KB
/
jose_deg_and_enrichment_figures.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
###### Load packages ######
# Load necessary packages for single cell RNA-Seq analysis including packages for downstream Gene Ontology Analysis
library(monocle)
library(plyr)
library(stringr)
library(dplyr)
library(ggplot2)
library(reshape2)
library(piano)
library(tidyr)
library(viridis)
library(ggrepel)
source('helper_functions.R')
# Colors for plots
cluster_colors=c("1"="#e41a1c", "2"="#377eb8", "3"="#4daf4a", "4"="#984ea3", "5"="#ff7f00", "6"="#a65628", "NONE"="black")
# Load in data
informative_filelist <- Sys.glob("temp_data/pairwise_deg/informative/singles/cropseq/*.txt")
all_pairwise_informative_deg_results_gene = do.call(rbind, lapply(informative_filelist, read.delim))
all_pairwise_informative_deg_results_gene <- all_pairwise_informative_deg_results_gene[,-1]
dox_100nm_cds <- readRDS("temp_data/barcode_enrichment/dox_100nm_cds.rds")
# Preprocess RNA-seq data
dox_100nm_expressed_genes <- row.names(fData(dox_100nm_cds)[rowSums(exprs(dox_100nm_cds) > 0) > 50 ,])
dox_100nm_informative_cds <- dox_100nm_cds[,pData(dox_100nm_cds)$informative == TRUE]
preprocess_cds <- function(cds){
cds <- detectGenes(cds, min_expr = 0.5)
cds <- estimateSizeFactors(cds)
cds <- estimateDispersions(cds)
}
dox_100nm_informative_cds <- preprocess_cds(dox_100nm_informative_cds)
# Get the cluster-target associations where enrichment was found
dox.cluster_target_pairs = pData(dox_100nm_informative_cds) %>% filter(informative == TRUE & gene != "NONTARGETING") %>% select(gene, Cluster) %>% unique()
# Overlap between DEGs and TP53 set of DEGs for cells where their target-cluster pair was found to be enriched
dox_singles_deg_results <- all_pairwise_informative_deg_results_gene[all_pairwise_informative_deg_results_gene$column_value == "dox_100nm",]
dox_singles_significant_degs <- dox_singles_deg_results[dox_singles_deg_results$qval < 0.05,]
dox_singles_significant_genes = unique(dox_singles_significant_degs$id)
dox.tp53_genes = unique(subset(dox_singles_significant_degs, target == 'TP53')$id)
dox.tp53_deg_overlap = dox_singles_significant_degs %>%
group_by(target) %>%
summarize(overlap=sum(id %in% dox.tp53_genes), not_overlap=sum(! id %in% dox.tp53_genes), proportion=overlap / (overlap + not_overlap)) %>%
arrange(proportion) %>%
mutate(target = factor(target, levels=target))
dox.tp53_deg_overlap = merge(dox.tp53_deg_overlap, dox.cluster_target_pairs, by.x='target', by.y='gene')
## Dox
ggplot(dox.tp53_deg_overlap, aes(target, proportion)) +
geom_bar(stat='identity', color='black', aes(fill=Cluster)) +
theme_cfg() +
xlab('target') +
ylab('proportion\nTP53 DEG Overlap') +
ylim(0, 1) +
coord_flip() +
scale_fill_manual(values=cluster_colors, guide=FALSE) +
ggsave('supplemental_figures/dox.tp53_deg_overlaps.pdf', height=3.5, width=2.5)
# Now PC analysis
getMeanTargetExpression <- function(cds, DEGs){
gene_list <- sort(unique(pData(cds)$gene))
Mean_expression_byGene <- list()
Mean_expression <- list()
i <- 1
for (gene in gene_list){
cds_subset <- cds[DEGs,pData(cds)$gene == gene]
cds_subset_expr <- exprs(cds_subset)
cds_subset_expr <- Matrix::t(Matrix::t(cds_subset_expr)/sizeFactors(cds_subset))
cds_subset_Mean_expr <- as.matrix(rowMeans(cds_subset_expr), ncol = 1)
colnames(cds_subset_Mean_expr) <- as.factor(gene)
Mean_expression_byGene[[i]] <- cds_subset_Mean_expr
names(Mean_expression_byGene[[i]]) <- as.factor(gene)
i <- i + 1
}
Mean_expression <- do.call("cbind",Mean_expression_byGene)
return(Mean_expression)
}
dox_100nm_normalized_matrix <- getMeanTargetExpression(dox_100nm_informative_cds, dox_singles_significant_genes)
getScaledMatrix <- function(matrix){
scaled_matrix <- scale(t(scale(t(matrix))))
scaled_matrix[scaled_matrix > 3] <- 3
scaled_matrix[scaled_matrix < -3] <- -3
scaled_matrix[is.na(scaled_matrix)] <- 0
return(scaled_matrix)
}
dox_100nm_normalized_matrix_scaled = getScaledMatrix(dox_100nm_normalized_matrix)
dox_100nm_pca_matrix <- t(dox_100nm_normalized_matrix_scaled)
dox_pca <- prcomp(dox_100nm_pca_matrix, center = T, scale. = T)
dox_target_loadings <- data.frame(dox_pca$x,
.names = row.names(dox_pca$x))
dox_gene_loadings <- data.frame(dox_pca$rotation,
.names = row.names(dox_pca$rotation))
dox_target_loadings = merge(dox_target_loadings, dox.cluster_target_pairs, by.x="row.names", by.y="gene", all.x=T)
dox_target_loadings$Cluster = as.character(dox_target_loadings$Cluster)
dox_target_loadings$Cluster[is.na(dox_target_loadings$Cluster)] = "NONE"
dox_target_loadings$.names[dox_target_loadings$.names == 'NONTARGETING'] = 'NTC'
ggplot(dox_target_loadings, aes(x = PC1)) +
geom_histogram(color='black', aes(fill = Cluster)) +
theme_cfg() +
scale_fill_manual(values=cluster_colors, guide=FALSE) +
ggsave("supplemental_figures/dox.informative.pc1_histogram.png", width = 3.5, height = 1.75)
# Load necessary functions
source("jose_helper_functions/loadGSCSafe.R")
source('jose_helper_functions/plot_gsea_go.R')
source("jose_helper_functions/GSA_hyper_helper_functions.R")
## Load Gene Set Collections
hallmarksGSC <- loadGSCSafe(file="data/gmt_files/h.all.v6.0.symbols.gmt")
getGeneLoadingGSAhyperResults <- function(cds, expressed_genes, pca_res, gsc, PC_list, cutoff){
Ensembl_GSAlist <- as.matrix(fData(cds[expressed_genes])$gene_short_name)
rownames(Ensembl_GSAlist)<-row.names(fData(cds[expressed_genes]))
colnames(Ensembl_GSAlist) <- c("gene_short_name")
Ensembl_GSAlist<-Ensembl_GSAlist[,1]
Ensembl_GSAlist<-unique(toupper(Ensembl_GSAlist))
length(Ensembl_GSAlist)
GSAhyper.list <- list()
positive_gene_loadings_GSAhyper <- list()
negative_gene_loadings_GSAhyper <- list()
aload <- abs(pca_res$rotation)
sweep(aload, 2, colSums(aload), "/")
gene_loadings <- data.frame(pca_res$rotation,
.names = row.names(pca_res$rotation))
print(PC_list)
for(PC in PC_list){
top_positive_gene_loadings <- row.names(gene_loadings[gene_loadings[,PC] > cutoff[1],])
top_negative_gene_loadings <- row.names(gene_loadings[gene_loadings[,PC] < cutoff[2],])
print(length(top_positive_gene_loadings))
print(length(top_negative_gene_loadings))
positive_genes <- as.matrix(fData(cds[top_positive_gene_loadings])$gene_short_name)
rownames(positive_genes)<-row.names(fData(cds[top_positive_gene_loadings]))
colnames(positive_genes) <- c("gene_short_name")
positive_genes<-positive_genes[,1]
positive_genes<-unique(toupper(positive_genes))
length(positive_genes)
negative_genes <- as.matrix(fData(cds[top_negative_gene_loadings])$gene_short_name)
rownames(negative_genes)<-row.names(fData(cds[top_negative_gene_loadings]))
colnames(negative_genes) <- c("gene_short_name")
negative_genes<-negative_genes[,1]
negative_genes<-unique(toupper(negative_genes))
length(negative_genes)
positive_gene_loadings_GSAhyper[[PC]] <- collect_gsa_hyper_results_All(positive_genes,
Ensembl_GSAlist,
gsc)
negative_gene_loadings_GSAhyper[[PC]] <- collect_gsa_hyper_results_All(negative_genes,
Ensembl_GSAlist,
gsc)
}
GSAhyper.list[["positive_loadings"]] <- positive_gene_loadings_GSAhyper
GSAhyper.list[["negative_loadings"]] <- negative_gene_loadings_GSAhyper
return(GSAhyper.list)
}
Dox_geneLoadings_GSAhyper <- getGeneLoadingGSAhyperResults(dox_100nm_cds, dox_100nm_expressed_genes, dox_pca,
hallmarksGSC, c("PC1"), c(0.02,-0.02))
gsea_plot <- function(GSAhyper_list, qval_cutoff, pattern, sample, gsc, PC){
GSAhyper_df.positive <- as.data.frame(GSAhyper_list[['positive_loadings']][[PC]]$diffExp$p.adj)
GSAhyper_df.positive$loading= 'positive'
GSAhyper_df.positive$gene_set = row.names(GSAhyper_df.positive)
colnames(GSAhyper_df.positive) <- c("qval", 'loading', 'gene_set')
GSAhyper_df.negative <- as.data.frame(GSAhyper_list[['negative_loadings']][[PC]]$diffExp$p.adj)
GSAhyper_df.negative$loading= 'negative'
GSAhyper_df.negative$gene_set = row.names(GSAhyper_df.negative)
colnames(GSAhyper_df.negative) <- c("qval", 'loading', 'gene_set')
GSAhyper_df = rbind(GSAhyper_df.positive, GSAhyper_df.negative)
if(is.null(pattern) == FALSE){
GSAhyper_df$gene_set <- str_replace(string = GSAhyper_df$gene_set, pattern = pattern, replace = "")
}
GSAhyper_df$gene_set = str_replace_all(GSAhyper_df$gene_set, '_', ' ')
GSAhyper_df_cutoff <- GSAhyper_df %>% filter(qval < qval_cutoff) %>% arrange(desc(qval)) %>%
mutate(gene_set = factor(gene_set, levels = gene_set))
ggplot(GSAhyper_df_cutoff, aes(x = gene_set, y = -log10(qval))) +
geom_bar(stat = "identity", color='black', aes(fill = loading)) +
coord_flip() +
facet_wrap(~loading, ncol=1, scales='free') +
xlab('gene set')
}
gsea_plot(Dox_geneLoadings_GSAhyper, 0.01, pattern = "HALLMARK_", sample = "Dox_100nm", gsc = "hallmarksGSC", PC='PC1') +
theme_cfg() +
scale_fill_manual(values=c('positive'="#53585F", 'negative'='#53585F'), guide=FALSE) +
ggsave('supplemental_figures/dox.pc1.loadings.gsea.pdf', height=3.5, width=4.5)