-
Notifications
You must be signed in to change notification settings - Fork 0
/
gene_expression.Rmd
181 lines (126 loc) · 5.14 KB
/
gene_expression.Rmd
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
---
title: "Gene expression analysis"
author: "Karen Chu"
date: "7/31/2019"
output: html_document
---
Gene expression analysis.
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
Libraries.
```{r libraries}
library(DESeq2)
library(data.table)
library(TxDb.Mmusculus.UCSC.mm10.knownGene)
library(TxDb.Hsapiens.UCSC.hg19.knownGene)
library(org.Mm.eg.db)
library(org.Hs.eg.db)
library(biomaRt)
library(ggplot2)
library("ggrepel")
library(data.table)
library(dplyr)
```
Working directory.
```{r folder}
#folder <- "/Users/chuk/mount/chuk/LT_vs_MPPs/data/"
folder <-"/Users/chuk/sshfs_mount/chuk/LT_vs_MPPs/data/"
```
Import data. Need to be RDS object in order for downstream stuff to work.
```{r data lsc}
setwd(folder)
reads <- readRDS("rnaseq_read_count_entrez_id.rds")
data.name <- "HSPC"
name.type <- "WT_vs_KO"
reads.subset <- reads [ ,grepl(name.type, colnames(reads))]
reads.subset <- as.data.frame( assay(reads.subset) )
```
Filter by fpkm >=1
```{r fpkm}
fpkm.mat <- fpkm( DESeqDataSet( reads, ~1 ) )
fpkm.greater.equal.to.one <- which( rowMeans(fpkm.mat) >= 1 )
reads.filtered <- reads.subset [ rownames(reads.subset) %in% names(fpkm.greater.equal.to.one), ]
```
Prepare data for DESeq2.
```{r prep DESeq2}
condition <- ifelse(grepl("WT", colnames(reads.filtered)), "WT", "KO")
coldata <- data.frame(Sample = as.factor(colnames(reads.filtered)),
condition = as.factor(condition) )
dds <- DESeqDataSetFromMatrix(
countData = as.matrix(reads.filtered),
colData = coldata,
design = ~condition)
dds$cell <- relevel(dds$condition, "KO") # intercept that DESeq2 calculates depends on the first factor. If "LSK" first, then results in log2foldchange(LSC/LSK). If "LSC" first, then results is log2foldchange(LSK/LSC).
```
Run DESeq2
```{r DESeq2}
dds <- DESeq(dds)
dds.vst <- varianceStabilizingTransformation(dds)
res <- results(dds)
```
Choose below: org.Mm.eg.db for mouse annotation, or org.Hs.eg.db for human.
```{r gene symbol}
library(annotate)
entrezid.to.genesymbol <- function(res) {
res$entrez.id <- rownames(res)
read.count.entrez.id <- as.character(rownames(res))
read.count.gene.symbol <- lookUp(read.count.entrez.id, 'org.Mm.eg.db', 'SYMBOL')
#read.count.gene.symbol <- lookUp(read.count.entrez.id, 'org.Hs.eg.db', 'SYMBOL')
read.count.gene.symbol <- unlist(read.count.gene.symbol)
res$gene <- read.count.gene.symbol
return(res)
}
reads.filtered.with.genesymbols <- entrezid.to.genesymbol(res)
reads.filtered.with.genesymbols <- reads.filtered.with.genesymbols [ !is.na(reads.filtered.with.genesymbols$padj), ]
write.csv(as.data.frame(reads.filtered.with.genesymbols), paste0(folder, data.name, "_", name.type, "_DESeq2Results.csv"))
```
Filter out genes with padj = NA.
After running code, found out none of the p-adj are NA, so "res" and "lsc.lsk.dds.with.genesymbols" are the same dataframes but "lsc.lsk.dds.with.genesymbols" has gene symbols while "res" doesn't.
```{r padj NA filter}
cat("# of genes before padj filter: ", nrow(res))
res <- res[!is.na(res$padj), ]
cat("# of genes after padj filter: ", nrow(res))
```
PCA.
```{r PCA}
pdf( paste0(folder, data.name, "_", name.type, "_PCA-KC.pdf"), 8, 6, useDingbats = F )
pcaData <- plotPCA(dds.vst, intgroup=c("condition"), returnData=TRUE)
percentVar <- round(100 * attr(pcaData, "percentVar"))
#ggplot(pcaData, aes(PC1, PC2, color=Sample, shape=cell)) +
ggplot(pcaData, aes(PC1, PC2, color=condition)) +
geom_point(size=3) +
xlab(paste0("PC1: ",percentVar[1],"% variance")) +
ylab(paste0("PC2: ",percentVar[2],"% variance")) +
coord_fixed() +
ggtitle(name.type)
dev.off()
```
MA plot.
```{r MA plot}
pdf( paste0(folder, data.name, "_", name.type, "_MA-KC.pdf"), 8, 6, useDingbats = F )
plotMA(res, ylim=c(-5,5))
dev.off()
```
Volcano plot.
```{r volcano}
padj.thres <- 0.05
res <- reads.filtered.with.genesymbols
volcano.input <- as.data.frame( res[ order( res$padj, decreasing = F ), ] )
volcano.input <- mutate(volcano.input, sig=ifelse(volcano.input$padj < padj.thres, "Sig", "Not Sig"))
pdf( paste0(folder, data.name, "_", name.type, "_volcano-KC.pdf"), 14, 10, useDingbats = F )
ggplot(data=volcano.input, aes(x=log2FoldChange, y=-log10(pvalue), colour= sig)) +
geom_point(alpha=1, size=4) +
theme_bw() +
theme(legend.position="none") +
geom_text_repel(data=volcano.input[1:30,], aes(label=gene), size = 5,box.padding = unit(0.5, "lines"), point.padding = unit(0.5, "lines"), color="black") +
geom_vline(xintercept=c(-1,1)) +
xlab("\nlog2(LTST / MPP)") + ylab("-log10(p)\n") +
ggtitle(paste0(data.name, "; padj <", padj.thres, ";\n# of sig genes: ", nrow( subset(volcano.input, sig=="Sig") ), "\n")) +
scale_color_manual( values = c( "Sig"='red3', "Not Sig" ='darkgray' ) ) +
scale_x_continuous(breaks = c( round(min(volcano.input$log2FoldChange)):round(max(volcano.input$log2FoldChange)) )) +
theme(plot.title = element_text(size=40)) +
theme(axis.text=element_text(size=30, color="black"), axis.title=element_text(size=30, color="black")) +
theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "black", size=1))
dev.off()
```