-
Notifications
You must be signed in to change notification settings - Fork 0
/
MS_Data preprocessing
257 lines (193 loc) · 9.51 KB
/
MS_Data preprocessing
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
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
##################################################################################################################
##### PRE-PROCESSING Poly MeRIP-seq DATA ####
##################################################################################################################
## By Mark Larance and Renhua Song
# Raw data are available at the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD017391.
######## THIS PORTION IS RUN by Mark Larance ######
MaxQuant version 1.6.3.4
Peptide and protein level identification were set to a false discovery rate (FDR) of 1% using the human Uniprot database
### THIS PORTION OF THIS SCRIPT IS RUN IN R USING A .R SCRIPT ###
#############################################################################
## LIBRARIES TO USE
library(ggfortify)
library(cluster)
library(DESeq2)
library(apeglm)
library(limma)
library(edgeR)
library(clusterProfiler)
library(org.Hs.eg.db)
# Input data-------------------------------------------
## Raw intensity
df <- read.delim("filtered_NP_EXP_log2FC_Pvalue_ND_M0_M1_v2.txt",comment="#")
## Protein Imput LFQ intensity
count=df[,c(58:66,73:75)]
## Protein ID
rownames(count)<-df$Protein.IDs
## Sample gene_name
colnames(count)<-c("mo1","mo2","mo3","M01","M02","M03","M11","M12","M13","M21","M22","M23")
## Sample information
sample=data.frame(
sample=c("mo1","mo2","mo3","M01","M02","M03","M11","M12","M13","M21","M22","M23"),
condition=c(rep("mo",3),rep("M0",3),rep("M1",3),rep("M2",3))
)
####Select Protiens containing at least 6 samples' value.
count_raw=df[,c(37:39,22:24,28:30,34:36)] ##Raw LFQ intensity
count_data<- filter(count, rowSums(is.na(count_raw)) < 6 )
##### Quality Check : missing values #####
figa=data.frame(Percentage=as.numeric(colSums(is.na(count_data))/nrow(count_data)),
Count=(colSums(is.na(count_data))
))
figa$name=c("mo_r1","mo_r2","mo_r3","M0_r1","M0_r2","M0_r3","M1_r1","M1_r2","M1_r3","M2_r1","M2_r2","M2_r3")
figa$name=factor(figa$name,levels = figa$name)
figa$condition=c("mo","mo","mo","M0","M0","M0","M1","M1","M1","M2","M2","M2")
figa$condition=factor(figa$condition,levels=c("mo","M0","M1","M2"))
## plot the percentarge and count of missing value per sample
ggplot(data = figa, aes(name,Count, group=condition, fill=condition)) +
geom_bar(stat="identity", position = "dodge")+ scale_fill_manual(values=c("red","orange","blue","green"))+
geom_text(aes(label = paste0(round(100*figa$Percentage,1),"%"), vjust = -0.2))+theme_classic(base_size = 15) + theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))
################## Quality Check : Impute value distribution ##########
## read Imput LFQ intensity and log transfered
count=log10(df[,c(58:66,73:75)]+1)
rownames(count)=df$Protein.IDs
colnames(count)=c("mo_r1","mo_r2","mo_r3","M0_r1","M0_r2","M0_r3","M1_r1","M1_r2","M1_r3","M2_r1","M2_r2","M2_r3")
count_melt=melt(count)
count_melt$condition=sapply(strsplit(as.character(count_melt$variable),"_"),'[',1)
count_melt$condition=factor(count_melt$condition,levels=c("mo","M0","M1","M2"))
## plot the data distribution
ggplot(count_melt, aes(x=factor(variable), y=value, fill=condition)) +ylab("log10(Impute Intensity Value)")+
geom_boxplot()+ scale_fill_manual(values=c("red","orange","blue","green"))+theme_classic(base_size = 15)+xlab("") + theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1),legend.position = "NA")
##### Quality Check : overlapping protiens #####
##Define a function drawing flower plot
flower_plot2 <- function(sample, value, start, a, b,
ellipse_col = rgb(135, 206, 235, 150, max = 255),
circle_col = rgb(0, 162, 214, max = 255),
circle_text_cex = 1, labels=labels) {
par( bty = "n", ann = F, xaxt = "n", yaxt = "n", mar = c(1,1,1,1))
plot(c(0,10),c(0,10),type="n")
n <- length(sample)
deg <- 360 / n
res <- lapply(1:n, function(t){
ellipse_col <- ellipse_col[t]
plotrix::draw.ellipse(x = 5 + cos((start + deg * (t - 1)) * pi / 180),
y = 5 + sin((start + deg * (t - 1)) * pi / 180),
col = ellipse_col,
border = ellipse_col,
a = a, b = b, angle = deg * (t - 1))
text(x = 5 + 2.5 * cos((start + deg * (t - 1)) * pi / 180),
y = 5 + 2.5 * sin((start + deg * (t - 1)) * pi / 180),
value[t]
)
if (deg * (t - 1) < 180 && deg * (t - 1) > 0 ) {
text(x = 5 + 3.3 * cos((start + deg * (t - 1)) * pi / 180),
y = 5 + 3.3 * sin((start + deg * (t - 1)) * pi / 180),
sample[t],
srt = deg * (t - 1) - start,
adj = 1,
cex = circle_text_cex
)
} else {
text(x = 5 + 3.3 * cos((start + deg * (t - 1)) * pi / 180),
y = 5 + 3.3 * sin((start + deg * (t - 1)) * pi / 180),
sample[t],
srt = deg * (t - 1) + start,
adj = 0,
cex = circle_text_cex
)
}
})
plotrix::draw.circle(x = 5, y = 5, r = 1.5, col = circle_col, border = circle_col )
# tune location by x and y.
text(x = 4.7, y = 5, labels=labels)
}
## plot the overlapping protein numbers
flower_plot2 (c("mo_r1", "mo_r2", "mo_r3", "M0_r1", "M0_r2", "M0_r3",
"M1_r1", "M1_r2", "M1_r3", "M2_r1", "M2_r2", "M2_r3"),
c(as.numeric(nrow(count1)-colSums(is.na(count1)))), 90, 0.5, 2, labels="Core items: 3962",
ellipse_col = c(rep("red",3),rep("orange",3),rep("blue",3),rep("green",3)),
circle_col = "white" )
## Remove Proteins with NA value
count1=count[complete.cases(count), ]
## Create the DGEList data class
y <- DGEList(counts = count1)
## Normalization
y <- calcNormFactors(y)
## Clustr samples
mds <- plotMDS(y,top = 100)
## Note on factor levels
toplot <- data.frame(PC1 = mds$x*100, PC2 = mds$y*100, condition = sample$condition)
toplot$condition=factor(toplot$condition,levels = c("mo","M0","M1","M2"))
## Principal Components Analysis plot
ggplot(toplot, aes(PC1, PC2, color=condition)) +
geom_point(size=6,shape=0)+scale_color_manual(values = c("red","orange","blue","green"))+
xlab(paste0("PC1: ",11,"% variance")) +
ylab(paste0("PC2: ",70,"% variance")) +
theme_bw(base_size = 12)+theme(legend.position = "top")
### create design
condition=sample$condition
design=model.matrix(~0+condition,data = y$samples)
colnames(design) <- gsub("condition", "", colnames(design))
## Estimate dispersions
y <- estimateDisp(y, design, robust=TRUE)
## test for DE Proteins
fit <- glmQLFit(y, design, robust=TRUE)
contr <- makeContrasts("M1 - mo", levels=design)
qlf <- glmQLFTest(fit, contrast=contr)
is.de <- decideTests(qlf, p.value=0.05,adjust.method = "BH",lfc = 2)
summary(is.de)
tt<-topTags(qlf,n=483)
dep=as.data.frame(tt$table)
res1=subset(dep,abs(dep$logFC)>1&dep$FDR<0.05)
#################### Replicability of processed data ########
## Define a function to draw a scatter plot for a pair of variables (samples) with density colors
plotFun <- function(x,y){
dns <- densCols(x,y);
points(x,y, col="black", pch=".",cex=4,panel.first=grid());
abline(a=0, b=1, col="red")
}
## get the transformed data
nc <- as.data.frame(cpm(y, normalized.lib.sizes=TRUE))
count.table2=subset(nc,rownames(nc)%in%rownames(res1))[,1:3]
colnames(count.table2)=factor(c("mo replicate1","mo replicate2","mo replicate3"),levels =c("mo replicate3","mo replicate2","mo replicate1") )
## Plot the scatter plot for a few pairs of variables selected at random
set.seed(123) # forces the random number generator to produce fixed results. Should generally not be used, except for the sake of demonstration with a particular selection.
pairs(log2(count.table2[,sample(ncol(count.table2), 3)] + 1),
panel=plotFun)
#################### Gene Ontology analysis of processed data ########
## read the differential Mass spec between mo and M0 ##
MS_M0vsmo=read.csv("MS_M0vsmo.csv")
rownames(MS_M0vsmo)=MS_M0vsmo[,1]
mo_M0_GO<- enrichGO(gene =MS_M0vsmo$Gene,
OrgDb = org.Hs.eg.db,
universe = rownames(vsd_count),
keyType = 'SYMBOL',
ont = "BP",
pAdjustMethod = "BH",
pvalueCutoff = 1,
qvalueCutoff = 1)
## read the differential Mass spec between M0 and M1 ##
MS_M1vsM0=read.csv("MS_M1vsM0.csv")
rownames(MS_M1vsM0)=MS_M1vsM0[,1]
M0_M1_GO<- enrichGO(gene = MS_M1vsM0$Gene,
OrgDb = org.Hs.eg.db,
universe = rownames(vsd_count),
keyType = 'SYMBOL',
ont = "BP",
pAdjustMethod = "BH",
pvalueCutoff = 1,
qvalueCutoff = 1)
## read the differential genes between M0 and M2 ##
MS_M2vsM0=read.csv("MS_M2vsM0.csv")
rownames(MS_M2vsM0)=MS_M2vsM0[,1]
M0_M2_GO<- enrichGO(gene = MS_M2vsM0$Gene,
OrgDb = org.Hs.eg.db,
universe = rownames(vsd_count),
keyType = 'SYMBOL',
ont = "BP",
pAdjustMethod = "BH",
pvalueCutoff = 1,
qvalueCutoff = 1)
### Convert to DataFrame #####
GO_mo_M0_MS=as.data.frame(mo_M0_GO)
GO_M0_M1_MS=as.data.frame(M0_M1_GO)
GO_M0_M2_MS=as.data.frame(M0_M2_GO)