-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy patho-boxplots.R
executable file
·163 lines (133 loc) · 6.45 KB
/
o-boxplots.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
#!/usr/bin/env Rscript
suppressPackageStartupMessages(library(optparse))
suppressPackageStartupMessages(library(ggplot2))
suppressPackageStartupMessages(library(MASS))
suppressPackageStartupMessages(library(gtools))
suppressPackageStartupMessages(library(compute.es))
# command line options
option_list <- list(
make_option(c("--mapping_var"),
help = "Mapping variable to use from the sample mapping file. If none specified, the first column is used."),
make_option(c("--output_directory"), default=".",
help = "Output directory to store images [default \"%default\"]"),
make_option(c("--remove_outliers"), default=F,
help = "Remove upper outliers for better scaling [default \"%default\"]"),
make_option(c("--pdf_height"), default=9,
help = "PDF output height [default \"%default\"]"))
parser <- OptionParser(usage = "this_script.R [options] input_matrix sample_mapping", option_list=option_list,
description="Visualize the distribution of a unit among mapping categories")
arguments <- parse_args(parser, positional_arguments = TRUE)
options <- arguments$options
# check if the positional argument is set
if(length(arguments$args) != 2) {
cat("Incorrect number of required positional arguments\n\n")
print_help(parser)
stop()
}
input_matrix <- arguments$args[1]
sample_mapping <- arguments$args[2]
# check if the input file is accessible
if(file.access(input_matrix) == -1){
stop(sprintf("Input matrix '%s' does not exist", input_matrix))
}
if(file.access(sample_mapping) == -1){
stop(sprintf("Sample mapping '%s' does not exist", sample_mapping))
}
data <- as.data.frame(read.csv(input_matrix, header=TRUE, sep="\t"))
metadata <- as.data.frame(read.csv(sample_mapping, header=TRUE, sep="\t"))
metadata <- subset(metadata, metadata$samples %in% data$samples)
metadata$samples <- factor(metadata$samples)
metadata <- metadata[with(metadata, order(samples)),]
data <- data[with(data, order(samples)),]
if(invalid(options$mapping_var)){
mapping_variable <- names(metadata)[2]
} else {
if(options$mapping_var %in% names(metadata)){
mapping_variable <- options$mapping_var
} else {
stop(sprintf("Mapping variable '%s' does not exist in the mapping file", options$mapping_var))
}
}
pdf_width <- length(levels(metadata[[mapping_variable]]))
if(pdf_width < 4)
pdf_width <- 4
nodes <- names(data[, !names(data) %in% c("samples")])
sums <- rep(NA, length(nodes))
for(i in 1:length(nodes)){
sums[[i]] <- c(sum(data[[nodes[i]]]))
}
ord <- order(sums, decreasing = TRUE)
get_first_two_factors_and_effect_size <- function(d, mapping_variable, node){
#Êperform lda on the node with respect to classes
lda_result <- suppressWarnings(lda(d[[mapping_variable]] ~ ., data = data.frame(d[[node]]), tol=0.00001))
# find the two factors with the highes mean for this node
factors_means_df <- data.frame(levels(factor(d[[mapping_variable]])), as.numeric(lda_result$means))
colnames(factors_means_df) <- c('factors', 'means')
ordered_factors_means_df <- factors_means_df[with(factors_means_df, order(-means)), ]
factor_1 <- as.character(ordered_factors_means_df[1,]$factors)
factor_2 <- as.character(ordered_factors_means_df[2,]$factors)
#Êlda scaling
w <- lda_result$scaling[,1]
scaled_frequencies <- as.matrix(d[[node]])%*%w
effect_size <- abs(mean(scaled_frequencies[d[[mapping_variable]]==factor_1]) - mean(scaled_frequencies[d[[mapping_variable]]==factor_2]))
return(data.frame(factor_1, factor_2, effect_size))
}
mapping <- with(metadata, get(mapping_variable))
mapping <- factor(mapping, levels=levels(mapping))
for (i in 1:length(nodes[ord])){
node <- nodes[ord][i]
#Êget the subset of data with the mapping and node of interest i = 1
d <- data.frame(data[[node]], metadata[[mapping_variable]])
names(d) <- c(node, mapping_variable)
mapping <- d[[mapping_variable]]
mapping <- factor(mapping, levels=levels(mapping))
if(options$remove_outliers == T){
upper_whisker <- boxplot.stats(d[[node]])$stats[c(1, 5)][2] * 2
outlier_limit <- mean(d[d[[node]] > upper_whisker, ][[node]])
if(!is.nan(outlier_limit)){
d <- d[d[[node]] < outlier_limit, ]
mapping <- d[[mapping_variable]]
mapping <- factor(mapping, levels=levels(mapping))
}
}
k <- kruskal.test(d[[node]], g = mapping)
kruskal_p <- k$p.value
kruskal_s <- as.numeric(k$statistic)
#if (kruskal_p > 0.05)
# next
lda_result <- get_first_two_factors_and_effect_size(d, mapping_variable, node)
factor_1 <- as.character(lda_result$factor_1)
factor_2 <- as.character(lda_result$factor_2)
lda_effect_size <- lda_result$effect_size
#Êfactor 1 is the highest mean. lets proceed with that.
d.test <- d[which(d[[mapping_variable]] %in% factor_1), ][[node]]
d.ctrl <- d[which(d[[mapping_variable]] %in% factor_2), ][[node]]
d.mes_results <- mes(mean(d.test), mean(d.ctrl), sd(d.test), sd(d.ctrl), length(d.test), length(d.ctrl))
mes_d <- as.numeric(d.mes_results$MeanDifference[1])
mes_d_var <- as.numeric(d.mes_results$MeanDifference[2])
mes_g <- as.numeric(d.mes_results$MeanDifference[3])
mes_g_var <- as.numeric(d.mes_results$MeanDifference[4])
w <- wilcox.test(d.test, d.ctrl)
wilcox_p <- w$p.value
wilcox_s <- w$statistic
#if (wilcox_p > 0.05)
# next
out_text <- paste(i, node, factor_1, signif(kruskal_s, digits = 3), signif(kruskal_p, digits = 3), signif(wilcox_p, digits = 3), signif(lda_effect_size, digits = 3), sep="\t")
cat(paste(out_text, '\n', sep=''))
label <- paste("Kruskal test\np: ", signif(kruskal_p, digits = 2), ', Chi-S: ', signif(kruskal_s, digits = 3), sep = "")
label <- paste(label, "\n\nWilcoxon test\np: ", signif(wilcox_p, digits = 2), ', W: ', signif(wilcox_s, digits = 3), sep = "")
label <- paste(label, "\n\nMeans To Effect Size -> ", factor_1, "\n", sep="")
label <- paste(label, "d: ", signif(mes_d, digits = 2), ", var(d): ", signif(mes_d_var, digits = 2), sep="")
g <- ggplot(data=d, aes(mapping, d[[node]], color=mapping))
g <- g + geom_boxplot()
g <- g + geom_jitter(position = position_jitter(width = .05, height = 0), size=4, alpha=0.7)
g <- g + theme(axis.text.x = element_text(angle = 90, size=12, colour='black'))
g <- g + theme(legend.position = 'none')
g <- g + labs(x='', y='Percent Abundance', title = node)
#g <- g + annotate("text", x=(length(levels(factor(mapping))) + 1) / 2, y=max(d[[node]]) * 0.8, label=label, size=4)
pdf_output <- paste(options$output_directory, '/', sprintf("%03d", i), '_', node, ".pdf",sep="")
pdf(pdf_output, width = pdf_width, height = 8, pointsize = 8)#, family='mono')
print(g)
dev.off()
sprintf("PDF: '%s'", pdf_output)
}