-
Notifications
You must be signed in to change notification settings - Fork 3
/
PlotStructure.R
143 lines (134 loc) · 5.06 KB
/
PlotStructure.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
# Functions ot read in structure_f files
# TO DO:
# include into plot_struc:
# 1. sure palette colors =< k ; and give guidance
# 2. Reverse plot order
# 3. Plot by Pop | clustering
# Evaluate & include other clustering methods
install.packages("reshape")
install.packages("ggplot2")
install.packages("seriation")
install.packages("sata.table")
library(reshape)
library(ggplot2)
library(seriation) # Different clsutering methods for ordering series
library(data.table)
read_structure_f <- function(structure_file){
i <- grep("Inferred ancestry",readLines(structure_file, warn=FALSE))
j <- grep("Estimated Allele",readLines(structure_file, warn=FALSE))
struc_q <- read.table(structure_file, skip = i+1, nrows = j-i-4, row.names=1) # data frame containg sturcture q data per K
# struc_q <- struc_q[1:10,] # temprorary subset for dev, remove for final
struc_q[4] <- NULL
return(struc_q)
}
#' @importFrom seriation seriate get_order
cluster_structure <- function(q_matrix, method = 1){
METHODS = c("HC", "MDS")
if (all((1:2) != method)) {
stop("The method selection is not valid")
}
k <- ncol(q_matrix) # Column of last K in structure file
d <- dist(as.matrix(q_matrix[,4:k])) # infer sort order by cluster analysis
if (method == 1) {
sort_order <- get_order(seriate(d, method="HC"))
}
else if (method == 2) {
sort_order <- get_order(seriate(d, method="MDS"))
}
return(sort_order)
}
#' @importFrom reshape melt
format_for_plot <- function(q_matrix){
q_matrix <- q_matrix[sort_order,]
a <- melt(q_matrix[1:nrow(q_matrix),], id=c("V2"), measure.vars = c(4:ncol(q_matrix)))
a$V2 <- factor(a$V2, levels = rev(unique(a$V2)))
a <- data.table(a)
#recaculate all proportions to sum to exactly one:
a[, sum := sum(value), by=list(V2)]
a[, proportion := value/sum]
#delete unneeded rows
a <- a[, c("value","sum") := NULL]
plot_data <- a
return(plot_data)
}
#' @importFrom ggplot2 ggplot
plot_struc <-function(plot_data, method = 1, palette = 1){
METHODS = c("full", "reduced", "stripped")
if (all((1:3) != palette)) {
stop("The palette selection is not valid")
}
if (palette == 1) {
pal = "Set1"
}
else if (palette == 2) {
pal = "Set3"
}
else if (palette == 3) {
pal = "Paired"
}
if (all((1:3) != method)) {
stop("The method selection is not valid")
}
if (method == 1) {
q_plot <- ggplot(data = plot_data, aes(plot_data$V2, proportion, fill = variable, width=1)) +
geom_bar(stat = "identity") +
ylab("Proportion") +
xlab( "Individual") +
scale_fill_brewer(palette=pal, name="K", labels=c(1:length(unique(plot_data$variable)))) +
theme(
# axis.text = element_blank(),
# axis.ticks = element_blank(),
axis.text.x = element_text(angle =90),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.border = element_blank(),
panel.background = element_blank())
}
else if (method == 2) {
q_plot <- ggplot(data = plot_data, aes(plot_data$V2, proportion, fill = variable, width=1)) +
geom_bar(stat = "identity") +
ylab("Proportion") +
xlab( "Individual") +
scale_fill_brewer(palette=pal, name="K", labels=c(1:length(unique(plot_data$variable)))) +
theme(
# axis.text = element_blank(),
# axis.ticks = element_blank(),
axis.text.x = element_text(angle =90),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.border = element_blank(),
panel.background = element_blank())
}
else if (method == 3) {
q_plot <- ggplot(data = plot_data, aes(plot_data$V2, proportion, fill = variable, width=1)) +
geom_bar(stat = "identity") +
scale_fill_brewer(palette=pal, name="K", labels=c(1:length(unique(plot_data$variable)))) +
theme(
axis.text = element_blank(),
axis.ticks = element_blank(),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.border = element_blank(),
panel.background = element_blank())
}
return(q_plot)
}
structure_file <- "casia_f" # Read STRUCTURE file in xxx_f format and extract data for plotting
structure_file <- "MX_two_run_40_f" # Read STRUCTURE file in xxx_f format and extract data for plotting
structure_file <- "SA_New_run_19_f" # Read STRUCTURE file in xxx_f format and extract data for plotting
q_matrix <- read_structure_f(structure_file)
sort_order <- cluster_structure(q_matrix, method = 1)
plot_data <- format_for_plot(q_matrix)
plot_struc(plot_data, method = 3, palette = 1)
###############################################################
# Nik's sandbox
q_plot <- ggplot(data = plot_data, aes(plot_data$V2, proportion, fill = variable, width=1)) +
geom_bar(stat = "identity") +
scale_fill_brewer(palette="Set1", name="K", labels=c(1:length(unique(plot_data$variable)))) +
theme(
axis.text = element_blank(),
axis.ticks = element_blank(),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.border = element_blank(),
panel.background = element_blank())