-
Notifications
You must be signed in to change notification settings - Fork 1
/
ranking_binning_per_dataset-RAT.R
273 lines (206 loc) · 12.9 KB
/
ranking_binning_per_dataset-RAT.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
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
library(ggplot2)
library(dplyr)
library(stats)
library(tidyr)
library(viridis)
library(grid)
library(RColorBrewer)
library(heatmap3)
library(stringr)
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("sva")
library(sva)
BiocManager::install("limma")
library(limma)
# Input: ppb (FOT) normalised intensity values from MaxQuant,
# Normal tissue samples only
# Rename coloumns to add PXDID and Sample number
#
setwd('/Users/ananth/Documents/MaxQuant_Bechmarking/Rat/')
PXDID <- "PXD016958"
tmp <- read.table(file = paste(PXDID, "/proteinGroups_ppb_final-tissue_names_", PXDID, ".txt", sep=""), quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
#Some of the gene entries (ex. IGHA2 has two Ensembl gene ids ENSG00000211890 & ENSG00000276173
# ex. IGHV2-70 has two Ensembl gene ids ENSG00000274576 & ENSG00000282453)
# because of this there are duplicate entries of such genes. These are aggregated by taking the median of them
colnames(tmp)[2] <- "Gene.Symbol"
dataset <- aggregate(tmp[,-c(1)], list(tmp$Gene.Symbol), median, na.rm =TRUE)
dataset <- subset(dataset, select = -c(Gene.Symbol))
colnames(dataset)[1] <- "Gene"
dataset[dataset==0] <- NA
colnames(dataset)
colnames(dataset) <- gsub(".*_", "", colnames(dataset), perl=TRUE)
gene_names <- dataset[,c(1)]
convert_to_scaled_data <- function(data, bins) {
apply(data, 1, function(sample) {
ceiling((sample / max(sample, na.rm = T)) * bins)
})
}
convert_to_ranked_data <- function(data) {
apply(data, 2, function(sample_data){
unique_vals = unique(sample_data)
unique_vals = unique_vals[order(unique_vals)]
sapply(sample_data, function(s){which(s == unique_vals)}) - 1
})
}
Binning <- function(data_input, gene_names) {
# Make a data frame of sample + protein/peptide abundances.
data = data.frame(data_input)
# Rank data.
data_ranked = as.data.frame(t(convert_to_ranked_data(data)))
colnames(data_ranked) <- gene_names
# Different size of bins to try.
binning_size_approaches = list(
bins_2 = 1,
bins_5 = 4,
bins_10 = 9,
bins_100 = 99,
bins_1000 = 999,
bins_10000 = 9999
)
# Transform data using different bin sizes.
results = lapply(binning_size_approaches, function(binning_size_approach) {
# 'Bin' data by scaling the ranked data.
binned_data = as.data.frame(t(convert_to_scaled_data(data_ranked, binning_size_approach)))
})
}
# If a dataset has different tissues or tissue regions, do not bin all together,
# these tissues/regions must be separated and binned separately.
colnames(dataset)
#binning_input <- dataset[,-c(1)]
binning_input <- dataset[,c(31:33), drop=FALSE]
#backup before changing NAs to 0
binning_input_with_NA <- binning_input
# non-detected genes (NA or 0 values) are assigned 0 values here to help in binning.
binning_input[is.na(binning_input)] <- 0
Dataset_all_bins <- Binning(binning_input, gene_names)
# Consider only data grouped into 5 bins
Get_Bin_5 <- function(binned_data_list){
Bins_5 <- as.data.frame(t(binned_data_list$bins_5))
# rename bins 0 to 4 --> 1 to 5
# lowest intensities are in bin1 and highest in bin5
Bins_5[Bins_5 == 4] <- 5
Bins_5[Bins_5 == 3] <- 4
Bins_5[Bins_5 == 2] <- 3
Bins_5[Bins_5 == 1] <- 2
Bins_5[Bins_5 == 0] <- 1
Bins_5 <- tibble::rownames_to_column(Bins_5, "Gene")
}
Dataset_5_bins <- Get_Bin_5(Dataset_all_bins)
# Now reassign NA values back to those genes which were undetected or 0 in ALL samples
# NOTE: Only those genes are reassigned back to NA that have missing values in
# all samples (ie., if at least one sample had a value then it is left as 0)
# this particular reassigning is done after binning (later in the code).
Dataset_5_bins[which(apply(binning_input_with_NA, 1, function(x) all(is.na(x)) )),-c(1)] <- NA
write.table(Dataset_5_bins, file = paste("Binned_expression_", PXDID ,"_Third-segment-of-proximal-tubule_Kidney.txt", sep=""), sep = "\t", row.names = FALSE, quote = FALSE )
# Count the number of occurrences of each bin
Bins_count <- function(dataset){
Bin_1 <- apply(dataset, 1, function(x) length(which(x == 1)) )
Bin_2 <- apply(dataset, 1, function(x) length(which(x == 2)) )
Bin_3 <- apply(dataset, 1, function(x) length(which(x == 3)) )
Bin_4 <- apply(dataset, 1, function(x) length(which(x == 4)) )
Bin_5 <- apply(dataset, 1, function(x) length(which(x == 5)) )
result <- as.data.frame(cbind(Bin_1, Bin_2, Bin_3, Bin_4, Bin_5))
}
Bincounts <- cbind( Gene=Dataset_5_bins[,c(1)], as.data.frame( Bins_count ( as.data.frame(Dataset_5_bins[,-c(1)]) ) ) )
write.table(Bincounts, file = paste("Bin_counts_", PXDID, "_Third-segment-of-proximal-tubule_Kidney", sep=""), sep = "\t", row.names = FALSE, quote = FALSE )
# After this when comparing different datasets of the same tissue type, merge them
# this way there will be same genes between datasets when comparing
##### Comparing datasets
dataset1 <- read.table("Binned_expression_PXD001839_LeftVentricle.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset2 <- read.table("Binned_expression_PXD003375_Spinalcord-CaudalSegment.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset3 <- read.table("Binned_expression_PXD003375_Spinalcord-RostralSegment.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset4 <- read.table("Binned_expression_PXD004364_Testis.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset5 <- read.table("Binned_expression_PXD006692_Lung.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset6 <- read.table("Binned_expression_PXD012677_Amygdala.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset7 <- read.table("Binned_expression_PXD013543_Heart-LeftVentricle.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset8 <- read.table("Binned_expression_PXD015928_Tendon.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset9 <- read.table("Binned_expression_PXD016793_Liver.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset10 <- read.table("Binned_expression_PXD016958_Connecting-tubule_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset11 <- read.table("Binned_expression_PXD016958_Cortical-collecting-duct_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset12 <- read.table("Binned_expression_PXD016958_Cortical-thick-ascending-limb_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset13 <- read.table("Binned_expression_PXD016958_Distal-convoluted-tubule_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset14 <- read.table("Binned_expression_PXD016958_First-segment-of-proximal-tubule_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset15 <- read.table("Binned_expression_PXD016958_Inner-medullary-collecting-duct_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset16 <- read.table("Binned_expression_PXD016958_Medullary-thick-ascending-limb_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset17 <- read.table("Binned_expression_PXD016958_Outer-medullary-collecting-duct_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset18 <- read.table("Binned_expression_PXD016958_Second-segment-of-proximal-tubule_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
dataset19 <- read.table("Binned_expression_PXD016958_Third-segment-of-proximal-tubule_Kidney.txt" , quote = "\"", header = TRUE, sep = "\t", stringsAsFactors = FALSE, comment.char = "#")
# Keep the minimum number of genes that are common in all datsets
Merge_data <- function(dataset1, dataset2){
merged <- merge(x=dataset1, y=dataset2,
by.x=c("Gene"), by.y=c("Gene"), all.x=TRUE, all.y=TRUE)
}
merged_data <- Merge_data(dataset1, dataset2)
merged_data <- Merge_data(merged_data, dataset3)
merged_data <- Merge_data(merged_data, dataset4)
merged_data <- Merge_data(merged_data, dataset5)
merged_data <- Merge_data(merged_data, dataset6)
merged_data <- Merge_data(merged_data, dataset7)
merged_data <- Merge_data(merged_data, dataset8)
merged_data <- Merge_data(merged_data, dataset9)
merged_data <- Merge_data(merged_data, dataset10)
merged_data <- Merge_data(merged_data, dataset11)
merged_data <- Merge_data(merged_data, dataset12)
merged_data <- Merge_data(merged_data, dataset13)
merged_data <- Merge_data(merged_data, dataset14)
merged_data <- Merge_data(merged_data, dataset15)
merged_data <- Merge_data(merged_data, dataset16)
merged_data <- Merge_data(merged_data, dataset17)
merged_data <- Merge_data(merged_data, dataset18)
merged_data <- Merge_data(merged_data, dataset19)
filtered_data <- merged_data
# count number of samples for each gene that have non NA values (i.e., detected)
non_missing_sample_count_percentage <- data.frame(non_missing_sample_count_percentage= apply(filtered_data[2:ncol(filtered_data)], 1, function(x) (length(which(!is.na(x))))/(ncol(filtered_data)-1)*100 ))
filtered_data <- cbind(filtered_data, non_missing_sample_count_percentage)
# Filter genes that are present in at least 33%, 50% or 75% of samples
filtered_data <- filtered_data[filtered_data$non_missing_sample_count_percentage >= 50,]
### Since only common genes are considered, this condition is not required
# The undetected genes are also put into bin1
# (After discussing with Andy Jones [20/01/2021]: since undetected genes could be below detection threshold due to their low expression or abundance)
filtered_data[is.na(filtered_data)] <- 1
write.table(filtered_data, file = "Binned_filtered_data_non_missing_sample_count_percentage_greaterthanequalto_50.txt", sep = "\t", row.names = FALSE, quote = FALSE )
pca_input <- filtered_data[,-c(1, ncol(filtered_data))]
#### PCA plot
#
pca_bins <- prcomp(t(pca_input), scale = FALSE)
pca_plot_data <- data.frame(pca_bins$x[,1:2]) # Take components 1 and 2
pca_plot_data <- tibble::rownames_to_column(pca_plot_data, "Samples")
pca_plot_data$Organs <- gsub(".*\\.", "", pca_plot_data$Samples, perl=TRUE)
pca_plot_data$Organs <- gsub("Amygdala","Brain", pca_plot_data$Organs, perl=TRUE)
pca_plot_data$Organs <- gsub("LeftVentricle","Heart", pca_plot_data$Organs, perl=TRUE)
pca_plot_data$Organs <- gsub("CaudalSegment1|CaudalSegment2|CaudalSegment3|RostralSegment1|RostralSegment2|RostralSegment3","SpinalCord", pca_plot_data$Organs, ignore.case=T, perl=TRUE)
pca_plot_data$Organs <- gsub("Connectingtubule|Corticalcollectingduct|Corticalthickascendinglimb|Distalconvolutedtubule|Firstsegmentofproximaltubule|Innermedullarycollectingduct|Medullarythickascendinglimb|Outermedullarycollectingduct|Secondsegmentofproximaltubule|Thirdsegmentofproximaltubule","Kidney", pca_plot_data$Organs, ignore.case=T, perl=TRUE)
pca_plot_data <- pca_plot_data %>% group_by(Organs) %>% mutate( OrganID = cur_group_id() )
pca_plot_data$Datasets <- gsub("\\..*", "", pca_plot_data$Samples, perl=TRUE)
pca_plot_data <- pca_plot_data %>% group_by(Datasets) %>% mutate( DatasetID = cur_group_id() )
#Adding number of datasets next to each tissue sample
pca_plot_data <- pca_plot_data %>% group_by(Organs) %>% mutate( Organs = paste(Organs, "(", length(unique(Datasets)), ")") )
# To add lables on the legend
# From here: https://stackoverflow.com/questions/49965758/change-geom-texts-default-a-legend-to-label-string-itself
oldK <- GeomText$draw_key
# define new key
GeomText$draw_key <- function (data, params, size,
var=unique(pca_plot_data$DatasetID),
cols=scales::hue_pal()(length(var))) {
# sort as ggplot sorts these alphanumerically / or levels of factor
txt <- if(is.factor(var)) levels(var) else sort(var)
txt <- txt[match(data$colour, cols)]
textGrob(txt, 0.5, 0.5,
just="center",
gp = gpar(col = alpha(data$colour, data$alpha),
fontfamily = data$family,
fontface = data$fontface,
fontsize = data$size * .pt))
}
ggplot(pca_plot_data, aes(x=PC1, y = PC2, colour = Datasets))+
geom_text(aes(label=DatasetID))+
labs(x="PC1", y="PC2")+
theme_bw()+
theme(axis.text.x = element_text(angle = 90, hjust = 1))+
theme(legend.position = "bottom")+
theme(legend.key.size = unit(0.3,"line"))+
guides(col = guide_legend(ncol = 3))+
ggtitle("Rat-Samples bin_values_batch-per-dataset-regions\n[filter: genes detected in at least 50% of samples]")
# IMPORTANT reset key
GeomText$draw_key <- oldK