Moved to standard module architecture.

Added requirements, empty readme, and moved files to module directory

.gitignore

@@ -98,3 +98,4 @@ user.json
 *.pyc
 results_10/*
 output/*
+results/*

main.py → radseq_analyses_pipeline/main.py

 import os
 from analyse.analysis import analyse_directory
 
-root_dir = '/home/rferon/work/analyses/multi_species/rad_seq/denovo_map/scripts/pipeline/'
+root_dir = '/home/rferon/work/code/radseq_analyses_pipeline/'
 files_dir = os.path.join(root_dir, 'results_10')
 output_dir = os.path.join(root_dir, 'output')
 

visualise/frequencies.py → radseq_analyses_pipeline/visualise/frequencies.py

@@ -4,6 +4,6 @@ import os
 def frequencies(file, output_dir, n_individuals):
 
     scripts_d = ''.join(os.path.split(os.path.realpath(__file__))[:-1])
-    cmd = ('Rscript ' + os.path.join(scripts_d, 'ggplot.R') + ' ' + file + ' ' +
-           output_dir + ' ' + str(n_individuals))
+    cmd = ('Rscript ' + os.path.join(scripts_d, 'r_scripts', 'ggplot.R') + ' ' +
+           file + ' ' + output_dir)
     os.system(cmd)

visualise/ggplot.R → radseq_analyses_pipeline/visualise/r_scripts/frequencies.R

@@ -3,9 +3,9 @@ args = commandArgs(trailingOnly=TRUE)
 
 # test if there is at least one argument: if not, return an error
 if (length(args)==0) {
-    stop("At least one argument must be supplied (input file).n", call.=FALSE)
-} else if (length(args) != 3){
-    stop("Usage: R ggplot.R input_file.tsv output_dir n_individuals")
+    stop("At least one argument must be supplied (input file).", call.=FALSE)
+} else if (length(args) != 2){
+    stop("Usage: R frequencies.R input_file.tsv output_dir")
 }
 
 suppressMessages(library(readr))
@@ -14,28 +14,21 @@ suppressMessages(library(svglite))
 suppressMessages(library(scales))
 
 file = args[1]
-name = strsplit(file, '/')[[1]][length(strsplit(file, '/')[[1]])]
 output_dir = args[2]
-n_max = strtoi(args[3])
 
-split = strsplit(name, '_')[[1]]
-species_name = paste(split[1], "_", split[2], sep="")
-png_name = paste(species_name, "_", split[4], ".png", sep='')
-data = suppressMessages(read_delim(file, "\t", escape_double = FALSE, col_names = FALSE, trim_ws = TRUE))
-species = rep(species_name, dim(data)[1])
-m_value = rep(split[4],dim(data)[1])
-colnames(data) = c("Stack_ID", "Frequency")
+png_name = paste("tags_distribution.png", sep='')
 
-data = subset(data, data$Frequency <= n_max)
+data = suppressMessages(read_delim(file, "\t", escape_double = FALSE, col_names = TRUE, trim_ws = TRUE))
+colnames(data) = c("Frequency", "Count")
 
-g = ggplot(data, aes(x=Frequency)) +
-    geom_bar(aes(y = (..count..)/sum(..count..)), colour="black", fill="#CCCCCC") +
+g = ggplot(data, aes(x=Frequency, y=Count/sum(Count))) +
+    geom_bar(stat="identity", colour="black", fill="#CCCCCC") +
     geom_vline(xintercept = mean(data$Frequency), colour = "red") +
     geom_vline(xintercept = median(data$Frequency), colour = "blue") +
     ggtitle(paste("Total tags: ", dim(data)[1], sep='')) + theme(plot.title = element_text(hjust = 0.5)) +
     xlab("Number of individuals in which a tag is present") + ylab("Frequency (%)") +
     scale_y_continuous(labels=percent)
 
-png(paste(output_dir, gsub(".tsv", "", name), ".png", sep=''), width=1600, height=800, res=100)
+png(paste(output_dir, png_name, sep=''), width=1600, height=1000, res=130)
 print(g)
 x = dev.off()

radseq_analyses_pipeline/visualise/r_scripts/haplotypes.R

+#!/usr/bin/env Rscript
+args = commandArgs(trailingOnly=TRUE)
+
+# test if there is at least one argument: if not, return an error
+if (length(args)==0) {
+    stop("At least one argument must be supplied (input file).n", call.=FALSE)
+} else if (length(args) != 3){
+    stop("Usage: R haplotypes.R input_file.tsv output_dir treshold")
+}
+
+suppressMessages(library(readr))
+suppressMessages(library(ggplot2))
+
+file = args[1]
+output_dir = args[2]
+
+file = "~/work/code/radseq_analyses_pipeline/output/haplotypes_data.tsv"
+output_dir = "~/work/code/radseq_analyses_pipeline/results/"
+
+png_name = paste("sex_variable_haplotypes.png", sep='')
+
+data <- suppressMessages(read_delim(file, "\t", col_names = TRUE, escape_double = FALSE, trim_ws = TRUE))
+names(data) = c('Locus', 'Sequence', 'Males', 'Females', 'Male_outliers', 'Female_outliers')
+
+threshold = 5
+
+males = data.frame(table(data$Males[which(data$Males > threshold)], data$Females[which(data$Males > threshold)]))
+names(males) = c('Males', 'Females', 'Count')
+
+females = data.frame(table(data$Males[which(data$Females > threshold)], data$Females[which(data$Females > threshold)]))
+names(females) = c('Males', 'Females', 'Count')
+
+g <- ggplot(males, aes(x = Males, y = Count, fill = Females))
+g = g + geom_bar(stat='identity', position=position_dodge(), colour='black')
+
+h <- ggplot(females, aes(x = Females, y = Count, fill = Males))
+h = h + geom_bar(stat='identity', position=position_dodge(), colour='black')
+
+grid.arrange(g, h, ncol=2)

requirements.txt

+progress>=1.3
\ No newline at end of file